Manuscript Details

Manuscript number MICROC_2018_1186

Title Experimental design methodology for optimization and robustness determination

in ion pair RP-HPLC method development: Application for the simultaneous
determination of metformin hydrochloride, alogliptin benzoate and repaglinide in
tablets

Abstract
Experimental design methodology (DOE) surpasses traditional (one variable at a time) approach in improving the
chromatographic separation performance with minimum effort and resources. Since combination therapy of oral
hypoglycemic drugs was generally recommended by Clinical Practice Guidelines on hypoglycemic agent therapy for
treatment of type 2 diabetes mellitus, DOE was adopted in development of an ion pair RP-HPLC method for the
simultaneous determination of metformin hydrochloride (MET), alogliptin benzoate (ALO) and repaglinide (REP) in
combined binary tablets. Screening using Plackett-Burman design followed by face centered composite design
enabled the estimation of optimum settings that accomplish the most appropriate resolution and adequate peak shape
within suitable run time. The designed models were statistically analyzed, graphically presented by surface plots and
the relationships between coefficients of the derived polynomial equations were interpreted. The models fitted data
adequately and could be used for response prediction. Chromatographic separation was achieved using Inertsil ODS
column (250 mm, 4.6 mm, 5 µm), acetonitrile: phosphate buffer (0.01 M, adjusted to pH 2.5 with o-phosphoric acid):
0.3 % sodium heptane sulfonate in water (60: 20: 20, v/v/v) as a mobile phase at flow rate 1 mL min-1. UV detection
was carried out at 220 nm. Sharp and well separated peaks for the cited drugs were obtained. Laboratory prepared
mixtures containing MET, ALO and REP were analyzed with mean percentage recoveries 99.66 ± 0.468, 99.98 ±
0.398 and 99.70 ± 0.988, respectively. The method was successfully applied for the determination of the drugs in
binary tablets with good recoveries of 99.75 ± 0.6 and 99.74 ± 0.982 for MET and ALO tablets and 99.76 ±0.619 and
100.08 ± 1.159 for MET and REP tablets. Method validation was performed according to ICH guidelines Q2 (R1).
Robustness was evaluated using the created models. The developed method proved to be accurate, selective and
precise for the determination of the cited drugs in binary tablets and could be applied in quality control laboratories for
routine analysis of the drugs.

Submission Files Included in this PDF

File Name [File Type]
Highlights 1.docx [Highlights]

manuscript 4.docx [Manuscript File]

fig 1.jpg [Figure]

fig 2.jpg [Figure]

fig 3.jpg [Figure]

fig 4-1.jpg [Figure]

fig 4-2.jpg [Figure]

fig 5.jpg [Figure]

fig 6.jpg [Figure]

fig 7.jpg [Figure]

To view all the submission files, including those not included in the PDF, click on the manuscript title on your EVISE
Homepage, then click 'Download zip file'.
 RP-HPLC/UV simultaneous determination of metformin, alogliptin and repaglinide

 Screening of important factors affecting resolution using Plackett-Burman design

 Optimization of important factors using face centered composite design

 Statistical analysis of designed models for response prediction

 Method robustness was studied using the created models, thus saving time and cost

 Application to the analysis of cited drugs in binary tablets

1
2
3
Experimental design methodology for optimization and robustness determination in ion
4
5 pair RP-HPLC method development: Application for the simultaneous determination of
6
7 metformin hydrochloride, alogliptin benzoate and repaglinide in tablets
8
9
10
11 Marianne A. Mahrousea*, Nesrine T. Lamieb,c
12
13
14
15 aPharmaceutical Chemistry Department, Faculty of Pharmacy, Cairo University, Kasr El-
16
17 Aini Street, 11562, Cairo, Egypt
18 bPharmaceutical
19
Chemistry Department, Faculty of Pharmacy, King Abdulaziz University,
20 Jeddah
21
cAnalytical Chemistry Department, Faculty of Pharmacy, Cairo University, Kasr El-Aini
22
23 Street, 11562, Cairo, Egypt
24
25
26
* Corresponding author:
27
28 e-mail address: mariannealphonse@yahoo.com
29
30 orcid.org/0000-0003-0397-7952
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 1
59
60
61
62
Abstract
63
64 Experimental design methodology (DOE) surpasses traditional (one variable at a time)
65
66 approach in improving the chromatographic separation performance with minimum effort and
67 resources. Since combination therapy of oral hypoglycemic drugs was generally
68
69 recommended by Clinical Practice Guidelines on hypoglycemic agent therapy for treatment of
70
type 2 diabetes mellitus, DOE was adopted in development of an ion pair RP-HPLC method
71
72 for the simultaneous determination of metformin hydrochloride (MET), alogliptin benzoate
73
74
(ALO) and repaglinide (REP) in combined binary tablets. Screening using Plackett-Burman
75 design followed by face centered composite design enabled the estimation of optimum
76
77 settings that accomplish the most appropriate resolution and adequate peak shape within
78
suitable run time. The designed models were statistically analyzed, graphically presented by
79
80 surface plots and the relationships between coefficients of the derived polynomial equations
81
82
were interpreted. The models fitted data adequately and could be used for response prediction.
83 Chromatographic separation was achieved using Inertsil ODS column (250 mm, 4.6 mm, 5
84
85 µm), acetonitrile: phosphate buffer (0.01 M, adjusted to pH 2.5 with o-phosphoric acid): 0.3
86 % sodium heptane sulfonate in water (60: 20: 20, v/v/v) as a mobile phase at flow rate 1 mL
87
88 min-1. UV detection was carried out at 220 nm. Sharp and well separated peaks for the cited
89
drugs were obtained. Laboratory prepared mixtures containing MET, ALO and REP were
90
91 analyzed with mean percentage recoveries 99.66 ± 0.468, 99.98 ± 0.398 and 99.70 ± 0.988,
92
93 respectively. The method was successfully applied for the determination of the drugs in
94 binary tablets with good recoveries of 99.75 ± 0.6 and 99.74 ± 0.982 for MET and ALO
95
96 tablets and 99.76 ±0.619 and 100.08 ± 1.159 for MET and REP tablets. Method validation
97
was performed according to ICH guidelines Q2 (R1). Robustness was evaluated using the
98
99 created models. The developed method proved to be accurate, selective and precise for the
100
101 determination of the cited drugs in binary tablets and could be applied in quality control
102 laboratories for routine analysis of the drugs.
103
104
105
Keywords: metformin hydrochloride; alogliptin benzoate; repaglinide; face centered design;
106
107 HPLC; ion pair.
108
109
110
111
112
113
114
115
116
117 2
118
119
120
121
1. Introduction
122
123 Good glycemic control was established as the best way to avoid chronic diabetes
124
125 complications [1]. Combination therapy of oral hypoglycemic drugs was generally
126 recommended by Clinical Practice Guidelines on hypoglycemic agent therapy for treatment of
127
128 type 2 diabetes mellitus [1–3]. The rationale of multi targeted drugs in combination therapy is
129
based on achieving more rapid and improved glycemic control, reducing side effects
130
131 compared to monotherapy at maximum doses, improving adherence to treatment in addition
132
133
to complementary mechanisms of action [3–6]. Metformin hydrochloride (MET, Fig. 1a)
134 represents the cornerstone of dual therapy and is used extensively in combination with several
135
136 classes of oral hypoglycemic drugs [3]. It is chemically designated as 3-
137
(diaminomethylidene)-1,1-dimethylguanidine hydrochloride. Combination of MET with
138
139 alogliptin benzoate (ALO, Fig. 1b) or with repaglinide (REP, Fig. 1c) was marketed to
140
141
maintain the desired therapeutic target in patients with type 2 diabetes mellitus [7]. ALO is a
142 selective dipeptidyl peptidase-4 inhibitor, chemically known as 2-[[6-[(3R)-3-aminopiperidin-
143
144 1-yl]-3-methyl-2,4-dioxopyrimidin-1-yl]methyl]benzonitrile; benzoic acid. REP, 2-ethoxy-4-
145 [2-[[(1S)-3-methyl-1-(2-piperidin-1-ylphenyl)butyl]amino]-2-oxoethyl] benzoic acid, is a
146
147 nonsulfonylurea insulin secretagogue belonging to the meglitinide class with hypoglycemic
148
activity. REP has a rapid onset and short duration of action [8].
149
150 Several chromatographic methods were reported in literature for the determination of ALO
151
152 such as HPLC methods either in presence of impurities [9] or in combination with
153 pioglitazone [10–12] and LC-MS/MS method [13] in pharmaceutical dosage form and in
154
155 biological fluids. ALO/MET combination was estimated using HPLC [14–19] and LC-
156
MS/MS [20]. An LC-MS/MS method was described for the determination of MET and four
157
158 gliptins as linagliptin, sitagliptin and vildagliptin in plasma [21]. REP was determined by
159
160 HPLC in combination with MET [22–24] and other hypoglycemic drugs [25]. However no
161 methods were reported for the simultaneous determination of MET, ALO and REP in tablets.
162
163 Although some HPLC methods were described for the determination of ALO/MET,
164
REP/MET and MET in combination with other hypoglycemic drugs, experimental design and
165
166 the weights of each factor influencing the chromatographic separation has not been yet
167
168 statistically conducted. On the other hand, most of the reported chromatographic methods
169 were optimized employing the traditional procedure of one variable at a time (OVAT) which
170
171 was based on conducting a large number of experimental runs to find the most appropriate
172
resolution conditions [26,27]. To avoid consuming time and cost, design of experiment
173
174 (DOE) methodology was employed to facilitate the optimization step. DOE is an effective
175
176 3
177
178
179
180
technique that experimentally explores relations between studied experimental variables and
181
182 response. A minimal number of experimental runs represents a noticeable superiority over
183
184 OVAT approach [28,29].
185 According to the aforementioned aspects, the aim of this study was to apply DOE
186
187 methodology in developing an accurate and selective RP-HPLC method for the simultaneous
188
determination of MET, ALO and REP in combined dosage form. Two issues were taken into
189
190 consideration; first, method optimization using DOE in order to develop cost efficient method
191
192
while increasing its validity. The designed models for response prediction were statistically
193 analyzed, visualized by response surface plots and the relationships between coefficients of
194
195 the derived polynomial equations were interpreted. Furthermore, method robustness was
196
studied using the created models, thus saving time and cost. Second, implementation of ion
197
198 pair chromatography (IPC) to overcome the challenge of separating ionized drug molecules of
199
200
different pKa, IPC is an established and reliable technique that provides reduced separation
201 time with sharp peak shape and allows separation of ionized molecules of variable pKa.
202
203 Moreover, method validation was performed according to ICH guidelines Q2 (R1) [30].
204 2. Experimental
205
206 2.1. Instrument
207
Chromatographic system (Shimadzu instrument LC-10 AD VP) was equipped with a
208
209 vacuum degasser and UV/Vis detector. Separation was carried out on Inertsil ODS column
210
211 (250 mm, 4.6 mm, 5 µm). A sonicator (Power Sonic 405, Human Lab, Gwangju-si,
212 Gyeonggi, Korea) was employed. The pH measurements were adjusted using a pH meter
213
214 (Jenway, 3505, Essex, UK). Mobile phase filtration was carried out using membrane filters
215
(0.45 µm, Sartorius Stedim Biotech GmbH, Goettingen, Germany). The statistical software
216
217 Minitab® 17 (Minitab, Inc., State College, Pennsylvania, USA) was used in Experimental
218
219 design and data analysis [31].
220 2.2. Materials and Chemicals
221
222 MET and REP standards were generously obtained from Amoun Pharmaceutical
223
Company, Cairo, Egypt. Their purities were certified as 99.40 % and 99.6 %, respectively.
224
225 ALO standard was purchased from Shanxi Jinjin Chemical Co. ltd, Shanxi, China with purity
226
227 of 100.56 ± 0.945 [32]. Cidophage® 500 tablets (batch no. 3170094) were labeled to contain
228 500 mg MET, inhiglip® 12.5 tablets (batch no. 011) were labeled to contain 17 mg ALO
229
230 equivalent to alogliptin base 12.5 mg and novonorm® 2 mg tablets (batch no. 7010842) were
231
labeled to contain 2 mg REP. All tablets were purchased from local market. Bi-distilled water
232
233 was obtained using Aquatron Water Still (A4000D, Staffordshire, UK). Acetonitrile (HPLC
234
235 4
236
237
238
239
grade) and o-phosphoric acid were purchased from Sigma-Aldrich, Darmstadt, Germany.
240
241 Potassium dihydrogen phosphate (El-Nasr Pharmaceutical Chemicals Company, Egypt) was
242
243 used for buffer preparation. Sodium heptane sulfonate (HPLC grade) was purchased from
244 Fischer scientific, UK.
245
246 2.3. Stock Solutions
247
Standard stock solutions of MET (250 µg mL-1), ALO (170 µg mL-1) and REP (100 µg
248
249 mL-1) were prepared in methanol.
250
251
2.4. Sample Preparation
252 Twenty cidophage® 500 tablets, inhiglip® 12.5 tablets, novonorm® 2 mg tablets were
253
254 accurately weighed and ground separately to fine powder. Binary tablets formulation were
255
unavailable in Egyptian market, they were prepared in laboratory by mixing the powdered
256
257 tablets of (cidophage® 500 and inhiglip® 12.5 tablets) and (cidophage® 500 and novonorm® 2
258
259
mg tablets). Two quantities of the powdered tablets equivalent to (25 mg MET and 0.85 mg
260 ALO) and (25 mg MET and 0.1 mg REP) were transferred into two 50 mL volumetric flasks,
261
262 50 mL methanol were added and the flasks were sonicated for 15 min. After completing the
263 volume to the mark with methanol, sample stock solutions (500 µg mL-1 MET and 17 µg mL-1
264
265 ALO) and (500 µg mL-1 MET and 2 µg mL-1 REP) were obtained. The produced sample stock
266
solutions were filtered using Whatman filter paper. Aliquots from the resulting stock solutions
267
268 were introduced into two series of 10 mL volumetric flasks followed by dilution with the
269
270 mobile phase to the mark for the estimation of the studied drugs.
271 2.5. Experimental Design for Optimization of Chromatographic Conditions
272
273 2.5.1. Screening using Plackett-Burman design
274
Five factors were investigated at two levels; low (-1) and high (1). Eight experiments were
275
276 developed using Plackett-Burman design, performed in triplicate in random order to abolish
277
278 the effect of uncontrolled factors that may introduce bias into the measurements [33,34]. The
279 levels of the factors utilized in the screening design are shown in Table 1. Data were analyzed
280
281 utilizing the statistical software Minitab® 17 (Minitab, Inc., State College, Pennsylvania,
282
USA) [31] in order to find the most important factors.
283
284 2.5.2. Optimization using face centered composite design
285
286 Face centered composite (FCC) design was applied for optimization of the
287 chromatographic conditions. For two-level three factors, twenty experiments were designed
288
289 and run, as shown in Table 2 and the resolution of ALO peak from MET peak (ALO
290
Resolution) and the resolution of REP peak from ALO peak (REP Resolution) were
291
292 measured. The runs were conducted in random order to eliminate ascending or descending
293
294 5
295
296
297
298
299
forms in the residuals versus the order of gathering chromatographic data. For modeling of
300 the responses, two hierarchical second-order models were fitted using backward elimination,
301
302 including interaction terms. The models were employed to desgn response surface (3D) and
303
contour plots (2D) which allowed prediction of the optimal response. The analysis of variance
304
305 (ANOVA) was carried out to test for statistical relevance of the coefficients (at p < 0.05).
306
307 The model effectiveness was ensured using residual graphs and lack-of-fit test [35,36].
308 Optimum conditions for chromatographic separation of the studied drugs were established
309
310 using response optimizer tool.
311
2.6. Chromatographic Conditions
312
313 Chromatographic separation was performed using Inertsil ODS column (250 mm, 4.6 mm,
314
315 5 µm) as stationary phase. The mobile phase was comprised of acetonitrile: phosphate buffer
316 (0.01 M, adjusted to pH 2.5 with o-phosphoric acid): 0.3 % sodium heptane sulfonate in water
317
318 (60: 20: 20, v/v/v) at flow rate 1 mL min-1. UV detection was carried out at λ 220 nm.
319
Filtration of the mobile phase was performed through 0.45 m membrane filter followed by
320
321 degassing using ultrasonic bath. The chromatograph was equilibrated and saturated with the
322
323 mobile phase for half an hour before injecting the solutions. The analysis was conducted at
324 room temperature. Quantitation was performed based on peak area.
325
326 2.7. Procedure
327
2.7.1. Linearity
328
329 Accurately measured aliquots of standard stock solutions equivalent to 25-625 µg of MET,
330
331 17-425 µg of ALO and 1-400 µg of REP were separately transferred into three series of 10
332 mL volumetric flasks. The volume was completed to the mark with the mobile phase. An 20
333
334 µL aliquot of each solution was injected onto the column in triplicates. The previously
335
mentioned chromatographic conditions were achieved. Three calibration curves were
336
337 constructed by plotting area under the peak (AUP) against the corresponding drug
338
339 concentration (C).
340 2.7.2. Assay of Laboratory Prepared Mixtures
341
342 Various aliquots of (MET, ALO) and (MET, REP) standard stock solutions were
343
introduced into two series of 10 mL volumetric flasks and the volume wascompleted with the
344
345 mobile phase to obtain concentrations in the range of (30-62.5 µg mL-1 MET and 2.04-10.2 µg
346
347
mL-1 ALO) and (15-20 µg mL-1 MET and 0.1-10 µg mL-1 REP), respectively. The method
348 was conducted as stated under "Linearity" section and the concentrations of MET, ALO and
349
350 REP were calculated using the corresponding regression equations.
351
352
353 6
354
355
356
357
2.7.3. Assay of MET, ALO and REP in Bulk and in Pharmaceutical Preparation
358
359 The method mentioned under "Linearity" section was replicated employing concentrations
360
361 equivalent to 5-50 µg mL-1 MET, 3.4-38.25 µg mL-1 ALO and 0.3-35 µg mL-1 REP in bulk.
362 For the simultaneous estimation of binary tablets, dilution of the sample solutions was
363
364 performed followed by injection in triplicates. The concentrations of MET, ALO and REP
365
were calculated using the corresponding regression equations.
366
367 3. Results and Discussion
368
369
3.1. Experimental Design for Optimization of Chromatographic Conditions
370 DOE is a sequential process, employed to design and analyze experimental runs in order to
371
372 recognize the main factors and detect the factor settings that produce an optimum response.
373
One of the advantages of DOE is that at each step, only a few number of experimental runs
374
375 are conducted, thus saving time and costs. DOE was adopted including two types of designs;
376
377
screening design (fractional factorial design) to assess which of the factors were relevant and
378 important variables followed by the FCC design to create the response surface which assist in
379
380 identifying chromatographic factors that produce adequate resolution, good peak shape and
381 suitable retention time. [33]. Method optimization with minimum effort, time and resources
382
383 are the primary objectives of applying DOE in analytical method development [27,28,35].
384
3.1.1. Screening using Plackett-Burman design
385
386 At the beginning of experimentation, a large number of factors are expected to affect the
387
388 response. Screening designs are experimental designs that are required to detect the few most
389 important variables from a larger set of variables that have significant effect on the response
390
391 [33,34]. In this work, the factors expected to influence resolution and retention (based on
392
preliminary study) were % acetonitrile, buffer pH, % sodium heptane sulfonate, flow rate and
393
394 wavelength. Plackett-Burman design (fractional factorial design) was employed in this study,
395
396 eight experiments that included five factors were designed and experimentally run. Each
397 factor was investigated at two levels (-1, 1) as revealed in the design matrix, Table 1. As
398
399 shown in Fig. 2, the most important factors that greatly affect resolution of MET, ALO and
400
REP were % acetonitrile, buffer pH and % sodium heptane sulfonate. Therefore, these factors
401
402 were selected for further study in optimization design in order to achieve optimum
403
404 chromatographic conditions for resolution and retention of the studied drugs. The other two
405 factors were discarded as they did not seem to have an effect on the response, the flow rate
406
407 was fixed at 1 mL min-1 flow rate while UV detection at 220 nm.
408
409
410
411
412 7
413
414
415
416
3.1.2. Optimization using face centered composite design
417
418 As concluded from screening design, the important factors to be optimized were %
419
420 acetonitrile (X1), buffer pH (X2) and % sodium heptane sulfonate (X3). FCC design was built
421 using two-level three factors, the design matrix is shown in Table 2. Twenty experiments plus
422
423 six central points were conducted. The resolution was measured six times at the center point
424
(level zero) for each factor during the experiment in order to evaluate the experimental error.
425
426 The other experimental runs were conducted in random order with no repetition. The
427
428
measured responses that were considered were the resolution of ALO peak from MET peak
429 and the resolution of REP peak from ALO peak, hence two response surface models were
430
431 constructed and utilized to compute a non-linear relationship between significant factors and
432
response [26,27,35]. Based on the experimental results and regression analysis for FCC, the
433
434 second–order polynomial equations characterizing the quadratic models were computed using
435
436
the least square method and found to be:
437 Equation
𝐴𝐿𝑂(1) 1 2 3
2
2
2
3 1 2 135.1
1.896
𝑌1 =
43.29
3‒𝑋 2‒𝑋
328
33.119
‒𝑋110
+
438
439
440
441 Equation (2)
442
443 Where YALO and YREP are the resolution of ALO peak from MET peak and REP peak from
444
ALO peak, respectively, X1, X2 and X3 are % acetonitrile, buffer pH and % sodium heptane
445
446 sulfonate, respectively. X1X2, X1X3, X2X3 and X1X2X3 indicate the interaction between the
447
448
factors while X12, X22 and X32 are the quadratic terms of each factor. The response for any
449 probable factors levels can be predicted by simply replacing X1 X2 and X3, , without actually
450
451 running the experiments. The models take into consideration linear effects`, the interactions
452 between the investigated factors and quadratic effects. The other non-significant (p > 0.05)
453
454 parameters were not considered and were removed by backward elimination of terms
455
456
performed by Minitab® 17 statistical software [31].
457 The regression coefficient R2, presented in Table 3, gives a measure of the goodness of
458
459 fitting of the proposed equation and assist in estimation of model predictive power. High
460 values for R2 and adjusted R2 indicate that fitting of the data is adequate. In addition, high
461
462 predictive capability of the models for new measurements was ascertained depending on high
463
values of R2 predicted [28]. As shown in Table 3, the calculated p-value for lack-of-fit is
464
465 greater than 0.05; confirming that the models appropriately represent the experimental results,
466
467 at 95% confidence level.
468 3.1.3. Statistical Analysis of the Models
469
470
471 8
472
473
474
475
Statistical analysis of the models and significance of coefficients were computed by
476
477 applying ANOVA test employing the statistical software Minitab® 17 (Minitab, Inc., State
478
479 College, Pennsylvania, USA) [31]. As displayed in Table 4, the regression models are
480 significant (p = 0.000), revealing that the terms in the regression equations have a significant
481
482 effect on the response. The significant quadratic effects of % acetonitrile, buffer pH and %
483
sodium heptane sulfonate reveals that the relationship between these factors and resolution
484
485 follows a curved line rather than a straight line. The interaction plots, Fig. 3, illustrate relative
486
487
slopes of the lines, which confirm interaction between the factors. Since all the terms
488 significantly influence the resolution (p < 0.05) and the models can estimate response
489
490 effectively, they can be used for prediction of the resolution of ALO and REP [27,31].
491
3.1.4. Effects of the Factors
492
493 The second-order polynomial equations quantify the correlations between the measured
494
495
response Y and the experimental factors, based on regression analysis. The coefficients before
496 the terms of the equation supply a quantitative measure of the relevance of linear effects,
497
498 interactions between them and curvilinear effects of the variables.
499 Considering the linear terms, high value for the coefficient of linear term of % sodium
500
501 heptane sulfonate indicates that this factor dominate the other factors while buffer pH is more
502
significant than % acetonitrile in affecting the ALO and REP resolution. Negative value of the
503
504 linear terms reveals that ALO resolution is inversely related to % acetonitrile, buffer pH and
505
506 % sodium heptane sulfonate while REP resolution is inversely related to % sodium heptane
507 sulfonate and directly related to the other two factors. The individual effect of % sodium
508
509 heptane sulfonate is negative and at the same time its quadratic effect is positive, showing that
510
the resolution decreases as the factor level increases till a critical limit after which a more
511
512 increase leads to an increase of the response. Therefore, good resolution was obtained upon
513
514 using 0.3 % sodium heptane sulfonate.
515 An essential consideration can be concluded from the interactive terms, the high value of
516
517 coefficient of interaction term between buffer pH and % sodium heptane sulfonate indicates
518
that the effect of a change in the level of one factor varies depending on the value of the level
519
520 of the other factor. This correlation may be in a positive manner, i.e. to increase the response,
521
522 buffer pH is decreased while maintaining the % sodium heptane sulfonate at low level, for
523 ALO resolution or in a negative manner for REP resolution, i.e. to increase the response,
524
525 buffer pH is increased while keeping % sodium heptane sulfonate at low level. High values
526
for the coefficients of quadratic term of % sodium heptane sulfonate in the two models are
527
528 evidence of curvature in the response surface plot illustrating peak resolution. As a
529
530 9
531
532
533
534
conclusion, % sodium heptane sulfonate can relevantly influence the response of the model,
535
536 depending on the values of the respective factors.
537
538 Visualization of the response surface models assisted in describing precisely the extent by
539 which each chromatographic variable (% acetonitrile, buffer pH and % sodium heptane
540
541 sulfonate) affected the separation while interacting with other parameters. The polynomial
542
equations were graphically presented by response surface plots (three-dimensional plot) and
543
544 contour plots (two-dimensional plot) keeping one of the variables at the central point value,
545
546
Fig. 4. These plots reported the influence of interaction of each two factors on the response.
547 The contour plots are curved, revealing the non-linear effects of the factors on the response.
548
549 These figures permit the response prediction at any point of the experimental field [28,37].
550
In the presence of multiple response processes, the desirability function approach was
551
552 performed for their optimization. Optimum settings for separation were determined using the
553
554
response optimizer tool [34]. The mobile phase composed of acetonitrile: phosphate buffer
555 (0.01 M, adjusted to pH 2.5 with o-phosphoric acid): 0.3 % sodium heptane sulfonate in water
556
557 (60: 20: 20, v/v/v) was considered as the optimum mobile phase since appropriate resolution,
558 symmetric peak shape and suitable run time are obtained, Fig. 5.
559
560 3.1.5. Residual Analysis
561
Residual plots are used to evaluate the problems of non-normality, non-random variation
562
563 and outliers. As shown in Fig. 6, the errors are normally distributed with mean zero in the
564
565 normal probability plot of residuals, indicating absence of outliers. The residuals appear to be
566 randomly scattered about zero in residual plots versus fits and versus order, revealing that the
567
568 errors had constant variance and each error is independent of all other errors. Consequently,
569
residual analysis confirms that the models appropriately represent the data [28,35].
570
571 3.2. Ion pair chromatography
572
573 A problem was encountered in separation of the cited drug molecules of different pKa
574 values (pKa of MET 12.4, ALO 8.5 and REP 3.7). By changing the pH value of the mobile
575
576 phase, the charged drugs became non-ionized. However, resolution was reduced between
577
MET and ALO. IPC is better suited for ionized drug molecules of different pKa mixtures of
578
579 drugs. The mobile phase was supplemented with an ion-pair reagent (IP); sodium heptane
580
581 sulfonate. IP has a charge opposite to the drug as well as a hydrophobic region that allows
582 interaction with the stationary phase. By applying IPC technique, good resolution with sharp
583
584 peak was obtained with reduced separation time [38].
585
3.3. System Suitability Tests
586
587
588
589 10
590
591
592
593
System suitability test parameters involve column efficiency (number of theoretical plates),
594
595 resolution, capacity factor, selectivity factor, tailing of chromatographic peak, reproducibility
596
597 of retention time as percentage relative standard deviation (% RSD) and repeatability of peak
598 area for six injections of a solution of a 30 µg mL-1 as % RSD. Results displayed in Table 5.
599
600 assure that the reproducibility and resolution were adequate for the developed method [39].
601
3.4. Method Validation
602
603 The developed HPLC method was validated in accordance to ICH guidelines Q2 (R1) [30].
604
605
3.4.1. Linearity
606 Linearity was evaluated by analyzing six concentrations of each drug in tirplicates. High
607
608 values of regression coefficients confirmed good linearity between the peak areas and the
609
corresponding drug concentrations. The analytical data of the calibration curves comprising
610
611 standard deviations for the slope (Sb) and that of the intercept (Sa) are represented in Table 6.
612
613
3.4.2. Accuracy
614 To demonstrate the accuracy of the developed method, six different concentrations of
615
616 MET, ALO and REP were injected in triplicates, then percentage recoveries of the cited drugs
617 were calculated in bulk powder. The results, displayed in Table 6, indicate accuracy of the
618
619 method.
620
3.4.3. Precision
621
622 Intraday precision (repeatability) was ascertained by analyzing three concentrations of
623
624 MET (5, 37.5 and 50 μg mL-1), ALO (3.4, 17, 34 μg mL-1) and REP (1, 25, 35 μg mL-1) six
625 times. Repeatability was expressed in terms of % RSD which were found to be lower than
626
627 2% for the three concentrations. For interday precision, the aforementioned concentrations in
628
intraday precision were analyzed on three successive days to evaluate day to day ruggedness.
629
630 The results of intraday and interday precision are revealed in Table 6.
631
632 3.4.4. Selectivity
633 Selectivity assures the ability of the developed method to quantify the drug response in
634
635 presence of interferences. Selectivity was ascertained by analyzing MET, ALO and REP in
636
laboratory prepared mixtures containing the intact drugs at different ratios. Good resolution
637
638 between the drugs peaks, as revealed in Fig. 5, ensured method selectivity. Furthermore, the
639
640 chromatograms of (MET and ALO) and (MET and REP) in the sample solutions were
641 identical to those obtained by the drugs standard solutions and no extra peaks were detected,
642
643 Fig. 7. Furthermore, satisfactory results were obtained for the estimation of MET, ALO and
644
REP in dosage form, confirming the selectivity of the developed method, Table 6.
645
646 3.4.5. Limit of Detection and Limit of Quantitation
647
648 11
649
650
651
652
Limit of detection (LOD) and limit of quantitation (LOQ) were assessed using standard
653
654 deviation of the response and slope of the regression equation and results are displayed in
655
656 Table 6. Low values of LOD and LOQ reveal sensitivity of the developed method.
657 3.4.6. Robustness
658
659 Robustness of the developed method was investigated by alteration of some
660
chromatographic conditions such as % acetonitrile by ± 5 %, buffer pH by ± 0.1 and %
661
662 sodium heptane sulfonate by ± 0.02 %. The most important response to be investigated was
663
664
the resolution between the drug peaks. The resolution between MET and ALO peaks and
665 ALO and REP peaks were predicted using FCC design polynomial equations 1 and 2,
666
667 respectively. As displayed in Table 7, good values of the resolution, for all these alterations,
668
confirm good robustness of the developed RP-HPLC method.
669
670 3.5. Statistical analysis
671
672
The results obtained by the developed RP-HPLC method were statistically analyzed using
673 student’s t-test and variance ratio F-test at p = 0.05. The results are in good agreement with
674
675 those obtained using the reference method [17,24] for each drug, as revealed in Table 8,
676 indicating no significant difference between the methods in terms of accuracy and precision.
677
678 4. Conclusion
679
In the present study, a fruitful implementation of DOE in the optimization and robustness
680
681 estimation was adopted for the development of an ion-pair RP-HPLC method for the
682
683 simultaneous determination of MET, ALO and REP in tablets. Screening and FCC design
684 were employed to investigate and optimize the most important factors affecting
685
686 chromatographic resolution. The created response surface models were statistically validated
687
and used for response prediction. Regression analysis allowed the estimation of relationships
688
689 between the variables. Furthermore, IPC using sodium heptane sulfonate overcame the
690
691 challenge of separation of a mixture of drugs of different pKa. The developed method was
692 successfully applied for the determination of the studied drugs in binary tablets. Method
693
694 validation proved its accuracy, selectivity and precision. It be can be used for the routine
695
analysis of MET, ALO and REP in bulk powder and binary tablets in quality control
696
697 laboratories.
698
699 Conflicts of interest: The authors declare that they have no conflict of interest.
700
701
702 This research did not receive any specific grant from funding agencies in the public,
703
commercial, or not-for-profit sectors.
704
705
706
707 12
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761 References
762
[1] M.K. Moon, K.-Y. Hur, S.-H. Ko, S.-O. Park, B.-W. Lee, J.H. Kim, S.Y. Rhee, H.J.
763
764 Kim, K.M. Choi, N.-H. Kim, Combination Therapy of Oral Hypoglycemic Agents in
765
766 13
767
768
769
770
Patients with Type 2 Diabetes Mellitus, Diabetes Metab. J. 41 (2017) 357.
771
772 doi:10.4093/dmj.2017.41.5.357.
773
774 [2] Working group of the Clinical Practice Guideline for type 2 Diabetes. Clinical Practice
775 Guideline on type 2 Diabetes. Madrid: National Plan for the NHS of the MSC. Health
776
777 Technologies Assessment Agency of the Basque Country. Clinical Practice Guidelines
778
i, (2008).
779
780 [3] http://guidelines.diabetes.ca/docs/CPG-2018-full-EN.pdf, (n.d.).
781
782
[4] and M.G.R. C Rubina Reichal, Anti-diabetic Drugs and Fixed Dose Combination
783 Therapy, Res. Rev. J. Pharm. Pharm. Sci. 3 (2014) 55–58.
784
785 [5] M.D.H. Md Shahid Sarwar, Fixed Dose Combination and Disease Management, Int.
786
Res. J. Pharm. 3 (2012) 17–21.
787
788 [6] C.J. Bailey, C. Day, Fixed-dose single tablet antidiabetic combinations, Diabetes,
789
790
Obes. Metab. 11 (2009) 527–533. doi:10.1111/j.1463-1326.2008.00993.x.
791 [7] J. Neumiller, D. Holland, Alogliptin in combination with metformin and pioglitazone
792
793 for the treatment of type 2 diabetes mellitus, Diabetes, Metab. Syndr. Obes. Targets
794 Ther. (2014) 277. doi:10.2147/DMSO.S37648.
795
796 [8] R.G. Moses, Repaglinide/metformin fixed-dose combination to improve glycemic
797
control in patients with type 2 diabetes: an update., Diabetes. Metab. Syndr. Obes. 3
798
799 (2010) 145–54. http://www.ncbi.nlm.nih.gov/pubmed/21437084.
800
801 [9] K. Zhang, P. Ma, W. Jing, X. Zhang, A developed HPLC method for the determination
802 of Alogliptin Benzoate and its potential impurities in bulk drug and tablets, Asian J.
803
804 Pharm. Sci. 10 (2015) 152–158. doi:10.1016/j.ajps.2015.01.001.
805
[10] M.F. Abdel-Ghany, O. Abdel-Aziz, M.F. Ayad, M.M. Tadros, New LC–UV methods
806
807 for pharmaceutical analysis of novel anti-diabetic combinations, Acta Chromatogr. 29
808
809 (2017) 448–452. doi:10.1556/1326.2016.00089.
810 [11] A.A. El-Zaher, E.F. Elkady, H.M. Elwy, M.A.E.M. Saleh, Synchronized Determination
811
812 of Four Antidiabetic and Antihyperlipidemic Drugs by a Validated LC Method,
813
Chromatographia. 80 (2017) 87–97. doi:10.1007/s10337-016-3209-7.
814
815 [12] B. Haribabu, P.R. Krishna Veni, K.B. Murali Krishna, K.L. Prameela, Estimation of
816
817 Alogliptin and Pioglitazone Simultaneously in Combined Tablet Dosage Forms,
818 Marmara Pharm. J. 21 (2017) 345–345. doi:10.12991/marupj.300864.
819
820 [13] M.F. Abdel-Ghany, M.F. Ayad, M.M. Tadros, Enhanced LC–MS/MS analysis of
821
alogliptin and pioglitazone in human plasma: Applied to a preliminary pharmacokinetic
822
823 study, J. Chromatogr. B. 1058 (2017) 93–101. doi:10.1016/j.jchromb.2017.04.043.
824
825 14
826
827
828
829
[14] S. Mowaka, B.M. Ayoub, Comparative study between UHPLC-UV and UPLC-MS/MS
830
831 methods for determination of alogliptin and metformin in their pharmaceutical
832
833 combination., Pharmazie. 72 (2017) 67–72. doi:10.1691/ph.2017.6151.
834 [15] S.R.A. Chinnalalaiah Runja, P. Ravikumar, Stability indicating RP-HPLC method for
835
836 simultaneous estimation of alogliptin benzoate and metformin hydrochloride in tablet
837
dosage form, Int. J. Pharm. Pharm. Sci. 8 (2016) 116–120.
838
839 [16] M. chaitanya Koduru Swathi, Kalepu swathi, Method development for the
840
841
Simultaneous Estimation of Metformin and Alogliptin by using RP-HPLC, Int. J.
842 Pharma Res. Heal. Sci. 3 (2015) 747–753.
843
844 [17] B.M. Ayoub, O. Abdel-Aziz, A guide for using experimental design in
845
chromatographic method development: applied to the analysis of selected anti-diabetic
846
847 pharmaceutical combinations., Pharmazie. 71 (2016) 683–690.
848
849
doi:10.1691/ph.2016.6095.
850 [18] M.. R. Pavan kumar H.K, Manzoor Ahmed, Satishkumar Shetty A, Kuppast I.J,
851
852 Siddalingaswamy.M.S, RP-HPLC method development and validation for
853 simultaneous estimation of alogliptin benzoate and metformin hydrochloride in
854
855 combined tablet dosage form, World J. Pharm. Pharm. Sci. 4 (2015) 774–784.
856
[19] P.J.R. A. Praveen kumar, G. Aruna, K. Rajasekar, Analytical method development and
857
858 validation of alogliptin and metformin hydrochloride tablet dosage form by RP-HPLC
859
860 method, Int. Bull. Drug Res. 3 (2013) 58–68.
861 [20] S. Mowaka, E.F. Elkady, M.M. Elmazar, B.M. Ayoub, Enhanced LC-MS/MS
862
863 determination of alogliptin and metformin in plasma: Application to a pharmacokinetic
864
study, Microchem. J. 130 (2017) 360–365. doi:10.1016/j.microc.2016.10.002.
865
866 [21] M. Al Bratty, H.A. Alhazmi, S.A. Javed, K.G. Lalitha, M. Asmari, J. Wölker, S. El
867
868 Deeb, Development and Validation of LC–MS/MS Method for Simultaneous
869 Determination of Metformin and Four Gliptins in Human Plasma, Chromatographia. 80
870
871 (2017) 891–899. doi:10.1007/s10337-017-3288-0.
872
[22] L.K. Soni, T. Narsinghani, M. Jain, Development and validation of rp-hplc method for
873
874 simultaneous estimation of metformin hydrochloride and repaglinide in tablet dosage
875
876 form, J. Liq. Chromatogr. Relat. Technol. 35 (2012) 385–392.
877 doi:10.1080/10826076.2011.601492.
878
879 [23] S.S. Joshi, R.R. Nahire, N.R. Shastri, K. V. Surendranath, J. Satish, Validated stability-
880
indicating RP-HPLC UV method for simultaneous determination of metformin and
881
882 repaglinide, Acta Chromatogr. 24 (2012) 419–432. doi:10.1556/AChrom.24.2012.3.6.
883
884 15
885
886
887
888
[24] N.S.R. Manal M. Fouad, Development and validation of chromatographic and
889
890 spectroscopic methods for estimation of repaglinide and metfomin HCl in combined
891
892 dosage form, J. Glob. Trends Pharm. Sci. 5 (2014) 1844–1848.
893 [25] H.M.E. and M.A.S. Ehab F Elkady, Asmaa A El-Zaher, Validated Liquid
894
895 Chromatographic Method for Simultaneous Determination of Metformin, Pioglitazone,
896
Sitagliptin, Repaglinide, Glibenclamide and Gliclazide - Application for Counterfeit
897
898 Drug Analysis, J. Anal. Bioanal. Tech. (2015). doi:10.4172/2155-9872.S13-007.
899
900
[26] A. El-Zaher, M. Mahrouse, A validated spectrofluorimetric method for the
901 determination of nifuroxazide through coumarin formation using experimental design,
902
903 Chem. Cent. J. 7 (2013) 90. doi:10.1186/1752-153X-7-90.
904
[27] A.A. El-Zaher, M.A. Mahrouse, Utility of Experimental Design in Pre-Column
905
906 Derivatization for the Analysis of Tobramycin by HPLC—Fluorescence Detection:
907
908
Application to Ophthalmic Solution and Human Plasma, Anal. Chem. Insights. 8
909 (2013) ACI.S11612. doi:10.4137/ACI.S11612.
910
911 [28] B.A. Moussa, H.M.A. Hashem, M.A. Mahrouse, S.T. Mahmoud, Experimental Design
912 Approach in HPLC Method Development: Application for the Simultaneous
913
914 Determination of Sacubitril and Valsartan in Presence of Their Impurities and
915
Investigation of Degradation Kinetics, Chromatographia. 81 (2018) 139–156.
916
917 doi:10.1007/s10337-017-3425-9.
918
919 [29] M.J. Miller JN, Statistics and chemometrics for analytical chemistry, 6th ed., Pearson
920 education Limited, England, 2010.
921
922 [30] Commission of the European Communities (1996) Q2 (R1) Validation of analytical
923
procedures: Text and Methodology. In: Proceedings of the international conference on
924
925 harmonisation (ICH) of technical requirements for registration of pharmaceuticals for
926
927 human, (n.d.).
928 [31] Minitab®, Version 17, Statistical Software; Minitab, Inc.: State College, PA, USA,
929
930 2014. (2014). Version 17 edn., State College, PA, USA, (n.d.).
931
[32] A.K.S. D.B. Sen, A.K. Sen, A. Zanwar, R. Balaraman, Determination of alogliptin
932
933 benzoate and metformin hydrochloride in tablet dosage form by simultaneous equation
934
935 and absorption ratio method, Int. J. Pharm. Pharm. Sci. 7 (2015) 380–383.
936 [33] S.M.E.-A. Noura Hemdan Abou-Taleb, Dalia Rashad El-Wasseef, Dina Tawfik El-
937
938 Sherbiny, Application of Factorial Design for Optimization of Spectrophotometric
939
Determination of Cefdinir Using MBTH, Am. J. Pharmtech Res. 4 (2014) 238–258.
940
941 [34] R.S. Hanafi, M. Lämmerhofer, Response surface methodology for the determination of
942
943 16
944
945
946
947
the design space of enantiomeric separations on cinchona-based zwitterionic chiral
948
949 stationary phases by high performance liquid chromatography., J. Chromatogr. A. 1534
950
951 (2018) 55–63. doi:10.1016/j.chroma.2017.12.044.
952 [35] M. DC, Design and analysis of experiments, 7th ed., John Wiley & Sons, Inc., Asia,
953
954 2008.
955
[36] M.S. Deming SN, Experimental design: a chemometric approach, 2nd ed., Elsevier
956
957 Publishers, Amsterdam, London, New York, Tokyo, 1993.
958
959
[37] S.M. Khamanga, R.B. Walker, The use of experimental design in the development of
960 an HPLC–ECD method for the analysis of captopril, Talanta. 83 (2011) 1037–1049.
961
962 doi:10.1016/j.talanta.2010.11.025.
963
[38] https://www.sigmaaldrich.com/technical-documents/articles/analytix/ion-pair-
964
965 chromatography.html, (n.d.).
966
967
[39] The United States Pharmacopeia (USP 34), National Formulary (NF 29), 2011,
968 Rockville, MD, (n.d.).
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002 17
1003
1004
1005
1006
Table 1: Coded and actual values for the levels of the factors in Plackett-Burman screening
1007
1008 design
1009
1010
1011 Experimental variables
1012 Number of
1013experiments Coded values Actual values
1014
X1 X2 X3 X4 X5 % Acetonitrile Buffer pH % Sodium heptane Flow Wavelength
1015
sulfonate rate
1016
1017 1 -1 -1 -1 1 -1 50 2.5 0.1 1.5 220
1018
2 -1 1 1 -1 -1 50 7.0 0.3 0.5 220
1019
1020 3 -1 -1 1 -1 1 50 2.5 0.3 0.5 240
1021
4 -1 1 -1 1 1 50 7.0 0.1 1.5 240
1022
1023 5 1 1 -1 -1 -1 70 7.0 0.1 0.5 220
1024
6 1 -1 1 1 -1 70 2.5 0.3 1.5 220
1025
1026 7 1 -1 -1 -1 1 70 2.5 0.1 0.5 240
1027
8 1 1 1 1 1 70 7.0 0.3 1.5 240
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061 18
1062
1063
1064
1065
Table 2: Coded (Experimental matrix) and actual values (experimental plan) for the levels of
1066
1067 the factors in Face Centered Composite Design
1068
1069
1070
1071
Experimental variables
1072
1073 Number of
Coded values Actual values
1074 experiments
1075 % Sodium
X1 X2 X3 % Acetonitrile Buffer pH
1076 heptane sulfonate
1077 1 0 0 1 60 3.5 0.3
1078
1079 2 -1 -1 -1 50 2.5 0.1
1080 3 0 0 0 60 3.5 0.2
1081
1082 4 0 1 0 60 4.5 0.2
1083 5 0 -1 0 60 2.5 0.2
1084
1085 6 0 0 0 60 3.5 0.2
1086 7 1 1 -1 70 4.5 0.1
1087
1088 8 0 0 0 60 3.5 0.2
1089 9 0 0 -1 60 3.5 0.1
1090
1091 10 -1 1 -1 50 4.5 0.1
1092 11 1 0 0 70 3.5 0.2
1093
1094 12 0 0 0 60 3.5 0.2
1095 13 0 0 0 60 3.5 0.2
1096
1097 14 -1 0 0 50 3.5 0.2
1098 15 1 -1 1 70 2.5 0.3
1099
1100 16 0 1 1 50 4.5 0.3
1101 17 1 -1 -1 70 2.5 0.1
1102
1103 18 -1 -1 1 50 2.5 0.3
1104 19 1 1 1 70 4.5 0.3
1105
1106 20 0 0 0 60 3.5 0.2
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120 19
1121
1122
1123
1124
Table 3: Models fitting results
1125
1126
1127 Full quadratic models
1128 Model term
1129
ALO REP
1130 R2 99.54 99.72
1131
1132 Adjusted R2 99.02 99.24
1133
1134 Predicted R2 95.68 77.03
1135
1136 p-value for lack-of-fit 0.79 0.07
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179 20
1180
1181
1182
1183
Table 4: Analysis of variance (ANOVA) results for the models of ALO and REP
1184
1185 Full quadratic models
1186
1187 Source of variation ALO equation (1) REP Equation (2)
1188
1189 p- value p- value
1190 Model 0.000 0.000
1191
1192 Constant 0.000 0.000
1193
1194 % Acetonitrile 0.000 0.001
1195
1196 Buffer pH 0.000 0.000
1197
% Sodium heptane sulfonate 0.000 0.000
1198
1199
% Acetonitrile * % Acetonitrile ____ 0.007
1200
1201 Buffer pH * Buffer pH 0.009 0.000
1202
1203 % Sodium heptane sulfonate * % Sodium heptane sulfonate 0.000 0.000
1204
1205 % Acetonitrile * Buffer pH 0.000 0.000
1206
1207 % Acetonitrile * % Sodium heptane sulfonate 0.000 0.000
1208
Buffer pH * % Sodium heptane sulfonate 0.000 0.000
1209
1210 % Acetonitrile * % Acetonitrile * Buffer pH ____ 0.000
1211
1212 % Acetonitrile * % Acetonitrile * % Sodium heptane sulfonate ____ 0.000
1213
1214 % Acetonitrile * Buffer pH * Buffer pH 0.019 ____
1215
1216 % Acetonitrile * Buffer pH * % Sodium heptane sulfonate 0.000 0.000
1217
1218 Residual error
1219
Lack-of-fit 0.79 0.07
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238 21
1239
1240
1241
1242
Table 5: System suitability tests of the proposed RP-HPLC method for the simultaneous
1243
1244 determination of MET, ALO and REP
1245
1246
1247 Parameter MET ALO REP Reference value
1248 The higher the value, the
1249 N 361 3173 8190 more efficient the column is
1250
1251
R 3.74 10.18 >2
1252 T 0.75 0.90 1.07 ≤2
1253
1254 K´ 0.99 3.23 8.05 1 – 10
1255 α 3.58 2.50 ≥1
1256
% RSD of tR of 6
1257
injections
0.970 0.491 0.217
1258
% RSD of peak area of 6
1259
injections (concentration 0.162 0.102 0.112
1260
1261
30 µg mL-1)
1262
Where N: Number of theoretical plates, R: Resolution factor, T: Tailing factor, Kˊ: Capacity factor,
1263 α: Selectivity factor, tR Retention time.
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297 22
1298
1299
1300
1301
Table 6: Validation parameters and results obtained by the proposed RP-HPLC method for
1302
1303 the simultaneous determination of MET, ALO and REP
1304
1305
1306 Item MET ALO REP
1307
1308 Retention time (tR) (min) 1.925 4.192 8.733
1309 Wavelength of detection
1310 220 220 220
(nm)
1311
1312 Range of linearity (μg mL-1) 2.5 – 62.5 1.7 – 42.5 0.1 - 40
1313
1314 Slope 59.8362 19.9879 46.234
1315
1316 Intercept 49.7938 4.6577 0.0676
1317
1318 Regression coefficient (r2) 0.9995 0.9999 1
1319
1320 LOD* (μg mL-1) 0.391 0.383 0.027
1321
1322 LOQ* (μg mL-1) 1.185 1.161 0.081
1323 Standard deviation of the
1324 0.762 0.079 0.142
slope (Sb)
1325 Standard deviation of the
1326 26.613 1.875 3.182
intercept (Sa)
1327 Confidence limit of the
59.8362±2.425 19.9879±0.251 46.234±0.394
1328 slope
1329 Confidence limit of the
49.7938±84.683 4.6577±5.966 0.0676±8.833
1330 intercept
1331 Standard error of
35.5747 2.5069 5.1244
1332 estimation
1333
1334 Intraday precision** 0.305-0.231-0.680 1.616-0.692-0.691 0.434-0.711-1.246
1335
1336 Interday precision** .485-0.070-0.699 1.293-0.779-0.558 0.782-0.662-0.357
1337
1338 Drug in bulk 99.89 ± 0.280 99.98 ± 0.398 100.17 ± 0.943
1339
1340
Drug in dosage form 99.75 ± 0.600 99.74 ± 0.982 100.08 ± 1.159
1341
1342
Drug added 99.78 ± 0.740 99.32 ± 0.908 100.71 ± 0.864
1343 Drug in laboratory
99.66 ± 0.468 99.98 ± 0.398 99.70 ± 0.988
1344 prepared mixtures
1345 * Limits of detection and quantitation are determined via calculations: LOD = 3.3*SD/slope, LOQ =
1346 10*SD/slope
1347 ** The intraday (n = 3), average of three concentrations of (5, 37.5 and 50 μg mL-1) for MET, (3.4, 17, 34 μg
1348 mL-1) for ALO and (1, 25, 35 μg mL-1) for REP repeated three times within the day.
1349 The interday (n = 3), average of three concentrations of (5, 37.5 and 50 μg mL-1) for MET, (3.4, 17, 34 μg mL-1)
1350 for ALO and (1, 25, 35 μg mL-1) for REP repeated three times in three successive days.
1351
1352
1353
1354
1355
1356 23
1357
1358
1359
1360
Table 7: Robustness study of the proposed RP-HPLC method using face centered composite
1361
1362 design
1363
1364
1365
1366 % Sodium
% Acetonitrile Buffer pH
1367
Parameter Drugs Standard heptane sulfonate
1368
1369 55 65 2.4 2.6 0.28 0.32
1370
1371 MET/ALO 3.74 3.75 3.35 3.75 3.36 3.43 3.75
1372 Resolution
1373 ALO/REP 10.18 8.98 11.13 10.37 2.37 10.77 12.26
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415 24
1416
1417
1418
1419
Table 8: Statistical comparison between the proposed RP-HPLC method for the simultaneous
1420
1421 determination of MET, ALO and REP in drug substance and the reference methods
1422
1423
1424 MET ALO REP
1425 Statistical
term RP-HPLC Reference RP-HPLC Reference RP-HPLC Reference
1426
method method [17] method method [17] method method [24]
1427
1428
Mean 99.89 99.78 99.98 99.24 100.17 99.26
1429
1430 n 6 6 6 6 6 6
1431 V 0.08 0.30 0.16 0.58 0.89 0.35
1432
± SD 0.280 0.55 0.398 0.76 0.943 0.59
1433
1434 ± SE 0.114 0.225 0.162 0.310 0.385 0.241
1435 RSD 0.280 0.551 0.398 0.766 0.941 0.594
1436
t (2.23)* 0.44 2.11 2.00
1437
1438 F (5.05)* 3.86 3.65 2.55
1439
1440 *Figures in parentheses are the theoretical t and F values at (p = 0.05).
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474 25
1475
1476
1477
1478
Figure captions
1479
1480 Fig.1 Chemical structures of metformin hydrochloride (a), alogliptin benzoate (b) and
1481
1482 repaglinide (c)
1483 Fig.2 Screening of the effects of % acetonitrile, buffer pH, % sodium heptane sulfonate, flow
1484
1485 rate and wavelength on the resolution of alogliptin benzoate (ALO) peak from metformin
1486
peak (ALO Resolution) (a) and that of repaglinide (REP) peak from ALO peak (REP
1487
1488 Resolution) (b)
1489
1490
Fig.3 Interaction plots for the effects of buffer pH, % acetonitrile and % sodium heptane
1491 sulfonate on the resolution of alogliptin benzoate (ALO) peak from metformin peak (ALO
1492
1493 Resolution) (a) and that of repaglinide (REP) peak from ALO peak (REP Resolution) (b)
1494
Fig.4 Surface plots (a) and contour plots (b) showing the effects of buffer pH, % acetonitrile
1495
1496 and % sodium heptane sulfonate on the resolution of alogliptin benzoate (ALO) peak from
1497
1498
metformin peak (ALO Resolution) and that of repaglinide (REP) peak from ALO peak (REP
1499 Resolution)
1500
1501 Fig.5 HPLC chromatogram of a laboratory prepared mixture of metformin hydrochloride
1502 (MET) (37.5 µg mL-1), alogliptin benzoate (ALO) (42.5 µg mL-1) and repaglinide (REP) (25
1503
1504 µg mL-1)
1505
Fig.6 Normal probability plot, histogram, residuals versus fits and residuals versus order plots
1506
1507 for resolution of alogliptin benzoate (ALO) peak from metformin peak (ALO Resolution) (a)
1508
1509 and that of repaglinide (REP) peak from ALO peak (REP Resolution) (b)
1510 Fig.7 HPLC chromatogram of metformin hydrochloride (MET) (30 µg mL-1) and alogliptin
1511
1512 benzoate (ALO) (1.02 µg mL-1) in mixed tablets (Inhiglip® 12.5 and Cidophage® 500 tablets)
1513
(a) and metformin hydrochloride (MET) (30 µg mL-1) and repaglinide (REP) (0.12 µg mL-1)
1514
1515 in mixed tablets (Novonorm® 2 and Cidophage® 500 tablets) (b)
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533 26
1534