VALIDATION OF THE SPECTROPHOTOMETRIC PROCEDURE FOR DESLORATADINE ASSAY IN TABLETS APPLYING THE UNCERTAINTY CONCEPT OF THE STATE PHARMACOPOEIA OF UKRAINE

Aim. This work aimed to validate an assay procedure for desloratadine tablets by direct spectrophotometric method. Materials and methods. A pilot-scale batch of the pharmaceutical preparation Alerdez, film-coated tablets containing 5 mg of desloratadine, manufactured by PJSC SIC «Borshchahivskiy CPP», Ukraine, was used as an object of the study. A UV-Vis spectrophotometer Lambda 25 (Perkin Elmer), analytical balance Mettler Toledo XP 205DR, and class A volumetric apparatus were used in the study. Validation of the procedure was performed following the metrological approach of the State Pharmacopoeia of Ukraine (SPhU), whose requirements for the target uncertainty and bias, which rest on the risk assessment of making incorrect decisions on compliance (a confidence level of 95 %), were translated into criteria for all validation characteristics recommended by ICH. All calculations were made in normalised coordinates. The linearity, accuracy and precision (repeatability) were studied in a single experiment using nine different concentrations that uniformly covered the range of ±30 % from the nominal concentration of desloratadine. For validation of the procedure, an SPhU reference standard of desloratadine was used. Results. The experiment design and validation characteristics being tested were in full compliance with ICH Q2(R1) recommendations. All performance characteristics conformed to the criteria recommended by the SPhU. Requirements for the target uncertainty (1.6 %) and bias for any systematic source of variation (≤ 0.51 %, negligible in relation to 1.6 %) were established. The analytical procedure was specific – the absorbance from the placebo solution was insignificant (A % = 0.36). The procedure met the requirements for linearity, accuracy, and precision at the repeatability level. The residual standard deviation s0 was 0.34 (≤ 0.84); correlation index Rc was 0.9998 (≥ 0.9991); intercept а was 0.045 (less than its confidence interval ∆a = 1.14). The confidence interval for recovery ∆Z, which was used as a precision estimate, was 0.55 % (less than the target uncertainty). The mean recovery, which was used as an accuracy estimate, statistically insignificantly deviated from 100 % (|Zmean–100| = 0.022 %). The confidence interval for the intermediate precision ∆intra was 0.33 % (less than the target uncertainty). The developed analytical procedure was found to be robust. Conclusions. A spectrophotometric procedure suitable for the assay of desloratadine in film-coated tablets Alerdez with content limits of ±5 % was validated by the SPhU approach.

Pharmacology, Toxicology and Pharmaceutical Science results with criteria set in advance. Therefore, it is topical to assess the suitability of the analytical procedure intended for SP assay of desloratadine by the SPhU approach, i.e. based on the measurement uncertainty concept and risk assessment of making a wrong decision about the compliance with specifications.
The objective of this work is to validate a procedure for assay of desloratadine in film-coated tablets (tablet strength 5 mg, content limits of desloratadine ±5 %) produced by PJSC SIC «Borshchahivskiy CPP», Ukraine, following the requirements of the SPhU.

Materials and methods 2.1. Test object
A pilot-scale batch of Alerdez (Lot No. Е01), 5 mg desloratadine film-coated tablets produced by «Borshchahivskiy CPP», Ukraine, served as an object of the study. The content of desloratadine per tablet is specified 5 mg ± 5 % (from 4.75 mg to 5.25 mg). The weight of one tablet is about 105 mg.

Reference standards and reagents
A desloratadine reference standard of the State Pharmacopoeia of Ukraine suitable for SP assay was used (the assigned value: 99.7 %; target uncertainty: 0.5 %, expressed as a one-sided confidence interval for a reliability level of 95 %).
All reagents used in the study met the requirements of the Ph. Eur. The desloratadine stock solution (1 mg/mL) was prepared for the linearity and stability studies. It was also used for the preparation of model solutions by the gravimetric method.

Equipment
The spectrophotometers Lambda 25 and Lambda 35* equipped with a 1-cm cuvette (Perkin Elmer), analytical balances Mettler Toledo XP 205DR and Kern ABS 220-4*, pH-meter Metrohm, Class A volumetric pipettes and flasks, PTFE filter with pre-filter 0.45 mm cat. No. SYTG0602MNXX104 manufactured by MDI were used in the study.
*Note: these were used additionally in the inter-laboratory precision experiment.

An analytical procedure being validated
The text of the analytical procedure is given below. Test solution. Grind 20 tablets to a visually homogeneous mass. Dilute 105 mg of the obtained sample in 250 mL of 0.1 M hydrochloric acid, keeping it in an ultrasonic bath for 15 minutes, and filter.
Reference solution. Dissolve 40 mg of the desloratadine reference standard in 100 mL of 0.1 M hydrochloric acid. Dilute 5.0 mL of the obtained solution in 100 mL of 0.1 M hydrochloric acid.
Measure the absorbance at a wavelength of 282 nm. The length of the optical path: 10 mm. The compensation solution: 0.1 М hydrochloric acid. The content of desloratadine (X 2 ) per tablet (in mg) is calculated by the formula: where A 1 -absorbance of the test solution; A 0 -absorbance of the reference solution; m 0 -test portion of the desloratadine reference standard, in mg; m 1 -test portion of the sample obtained by tablet grinding, in mg; P -assigned value of desloratadine in the reference standard, in per cent; b -average mass of tablets, in mg.
For the nominal concentration of desloratadine (0.02 mg/mL), the absorbance is about 0.64 absorption units (AU).

Design of the validation experiment and criteria
The detailed explanation of the theoretical principles of the SPhU approach is given in the monograph [19].

The target uncertainty of analytical results
Under the SPhU approach [11], the target uncertainty for assays of finished drug products (maxD As ) should be insignificant compared to the width of two-sided specifications: where B -half-width of content limits; maxD As -target uncertainty.
Hereafter under an expanded uncertainty, we understand a 95 % one-sided confidence interval. The coefficient of 0.32 provides a 95 % reliability of making a correct decision on compliance [13,19].
Then, for desloratadine tablets as a finished drug product with the content limits of ±5 %, the requirement for insignificance takes the following form: max . %. D As = 1 6 (3)

Use of normalised coordinates
When using normalised coordinates, we assume that during routine analysis, a single reference solution is used for calibration, i. e. the calibration curve passes through the origin.
The normalised coordinates are defined as follows: where C i -concentration of the analyte in the i-th model solution being analysed; C st -concentration of the analyte in the reference solution; A i -analytical signal of the analyte for the i-th model solution being analysed; A st -analytical signal of the analyte for the reference solution.
The concentration of the analyte in the reference solution should be close to the nominal concentration.
In normalised coordinates, for the calibration curve that passes through the origin, Z is the recovery expressed in per cent.
All further calculations are carried out in the normalised coordinates.

Criteria for performance characteristics Specificity
According to the SPhU approach, bias caused by the additional absorbance of the placebo components (d placebo ) should be insignificant in relation to maxD As : For content limits of ±5 %, the requirement for the insignificance of d placebo is expressed as follows:

Range
By the ICH recommendations [3], the range of the linearity study in coordinates (X i , Y i ) should be not less than 80-120 % for assay and not less than 70-130 % for the content uniformity.
Since the procedure is intended to be used for the assay and content uniformity, the range of 70-130 % and the most stringent acceptance criteria as for the assay with content limits of ±5 % (maxD As = 1.6 %) are applied.

Linearity, accuracy, precision (repeatability)
The linearity, accuracy, and precision (at the repeatability level) are studied simultaneously for nine concentrations, evenly distributed across the range, which complies with the ICH recommendations for precision at the repeatability level. Actual concentrations that were studied are given in Table 1 in normalised coordinates.

Linearity -the requirement for the residual standard deviation s 0
The residual standard deviation s 0 characterises a dispersion of the points around the line. Since each point can be treated as an individual analytical result, s 0 in the normalised coordinates is RSD for the standard uncertainty of analytical results. Therefore, the confidence interval calculated from s o should not exceed maxD As , i. e.: where g is the number of model solutions.
For g = 9 and content limits of ± 5 %, the requirement takes the following form:

Linearity -the requirement for the correlation index R c
For the estimation of validation results, the correlation index (R c ) is used following the SPhU approach: where X i -concentration of the i-th model solution used for the linearity study, in per cent of the nominal concentration; g -number of the model solutions.
Due to the relative complexity of the mathematical expression of the Pearson coefficient, it is difficult to formulate the requirements connecting the values of r with the requirements for the uncertainty of the analysis results [20]. Yet it is easy to develop requirements for R c based on those for maxD As . Unlike the Pearson coefficient, which is only applicable to the linear regression, R c can be used for any regression.
Taking into account the requirements for s 0 , the requirements for R c can be expressed as follows: SD range is calculated from the actual concentrations of model solutions.

Linearity -the requirement for the intercept (а)
For the intercept (а), a two-level criterion is used. The statistically insignificant difference from zero: the value of a should be less than its confidence interval.
Therefore, for g = 9, statistical insignificance can be expressed as follows: where s a is the standard deviation of the intercept of the line (a) obtained by the least-squares method. Pharmacology, Toxicology and Pharmaceutical Science The practically insignificant difference from zero: the value of a should be practically insignificant for the intended use of the procedure if the bias due to non-passage of the calibration curve through the origin does not exceed 0.32×maxD As .
For the symmetric range of application of the procedure, the criterion of practical insignificance takes the following form: where Х min is the minimum limit of the procedure application range. The criterion of practical insignificance (11) is only used in the event when the criterion of statistical insignificance (10) is not met.
For the content limits of ±5 % and the range of 70-130 %, the requirement takes the following form: . .

Precision -estimation from the linearity data
For the estimation of the procedure precision from the linearity study, Z i values are used: .
. 95 1 18595 The following relationship must be fulfilled: For content limits of ±5 %, the requirement takes the following form: Accuracy -estimation from the linearity data Bias (d) of the analytical results is defined as follows: By the SPhU approach, bias should be insignificant for spectrophotometric assay. For accuracy estimation, a two-level criterion is used. Statistical insignificance: the value of δ should be statistically insignificantly different from zero: where g is the number of solutions used for the accuracy study from the linearity data. For g = 9 and content limits of ±5 %, the following requirement for bias should be met: Practical insignificance: if the failure to comply with the requirements for statistical insignificance is observed, the practical insignificance δ should be checked: For content limits of ±5 %, the requirement for the practical insignificance is expressed as follows: . .

Intermediate precision
For the intermediate precision, we use the approach based on the SPhU concept, which is described in detail in our paper [15]. Analyse the same sample of ground tablets on m ≥ 2 different days. Prepare n ≥ 5 of independent test solutions each day, and analyse them varying the factors that could affect precision as much as possible (different analysts, equipment, etc.). If all factors affecting precision are insignificant, all the results must belong to the same population. Therefore, for all obtained assay results (Z i ), the mean (Z intra ), standard deviation (SD Zintra %), and relative confidence interval (∆ intra %) are calculated: ∆ intra should not exceed Δ As , For content limits of ±5 %, the requirement (25) takes the following form:

Robustness -stability of solutions
Based on the SPhU concept, we propose the following approach to stability studying. Prepare a test solution according to the procedure. Prepare a reference solution with precisely known concentrations of analyte («zero» solutions). Use the gravimetric method for dilutions. Assay the solutions («zero» time). After a predetermined period (we propose to check stability in N = 5 h, 24 h and 48 h after «zero» solutions preparation), prepare a fresh reference solution in the same manner (N-th solution). At the specified time, analyse both of the «zero» solutions using a fresh reference solution. Express results in per cents in relation to «added» amount for «zero» reference solution and in relation to «found» amount for «zero» test solution (Z N ); the desloratadine content in «zero» solutions is taken as 100 %. Calculate differences in concentrations (100 % -Z N ). The solutions are stable as long as the difference is insignificant in relation to maxD As : Since the specific absorbance of degradation products may differ from one of the analytes upward and downward, the difference modulus is used.
For content limits of ±5 %, the requirement (26) is expressed as follows: This approach allows us to confirm the stability of the solutions for quite a long period of time, which may be critical for a laboratory when investigating the OOS situation. However, the necessity to prepare solutions with the precisely known concentrations may be seen as a disadvantage.

Robustness -the reliability of the analysis with respect to deliberate variations in method parameters
The solubility of desloratadine is increased in acidic water solutions in comparison with pure water. The UV spectrum of desloratadine is also affected by pH of water diluent [18]. Consequently, it is reasonable to investigate the impact of HCl concentration variation in the diluent on the reference solution extinction and desloratadine recovery from the test portion of ground tablets.
For the prognosis of the «worst case» of 0.1 N⋅HCL concentration variability in the routine analysis, the uncertainty of the laboratory procedure of 0.1 N⋅HCl preparation was estimated per the SPhU recommendations. It comprised of about 2 %. On that basis, according to the SPhU insignificance principle, an analytical procedure will be robust if the influence of HCl concentration variability in the diluent on the reference solution absorbance is still insignificant for the range of HCl concentration of ±2 %´3.2 = ±7 % from the nominal concentration of 0.1 N·HCl. Following the same principle, the impact of HCl concentration variability is insignificant if the absorbance for reference solutions prepared using 0.093 N·HCl and 0.107 N·HCl does not differ from the absorbance of the reference solution prepared using 0.1 N·HCl by more than 0.51 %. In the same way, the variability of HCl concentration in the diluent has an insignificant impact on the desloratadine recovery if the «found» value for any test solutions prepared using 0.093 N·HCl, 0.1 N·HCl and 0.107 N·HCl does not deflect by more than 0.51 from the «true» value that should be obtained with the nominal HCl concentration. It means that the difference between «found» values for any two solutions must not exceed 0.51 %´Ö2 = 0.74 %, because both experimental values vary independently and can have uncertainty that should not exceed 0.51.

Experimental results
UV spectra of the reference solution, test solution, model solution, and placebo solution are given in Fig. 1.   Fig. 1. Specificity study: the UV spectrum of (A) Placebo Solution, (B) Desloratadine Reference Solution, (C) Test Solution, (D) Model Solution.

Specificity
The test solution prepared from the ground tablets for the stability evaluation, the «100 %» model solution prepared for the linearity evaluation and the reference solution showed identical spectra. Pharmacology, Toxicology and Pharmaceutical Science The extinction of this solution was recalculated per the nominal concentration of the desloratadine as follows: Then two solutions of placebo were prepared and measured following the procedure. The results of the specificity evaluation are shown in Table 1. The validation requirements for the specificity of the procedure are met.

Linearity. Accuracy. Precision
Results of the analysis of model solutions are given in Table 2. For the obtained results, the parameters of the linear relationship were calculated by the least-squares method: where Y i -concentrations of the model solutions, in normalised coordinates; X i -absorbance of the model solutions, in normalised coordinates; a -intercept; b -slope. The graph of the linear relationship is shown in Fig. 2.  Fig. 2. The linear relationship between normalised concentrations (X ) and normalised absorbance (Y ) for desloratadine model solutions from the linearity study

Linearity
Results of the calculation of parameters of the linear relationship between normalised concentrations (X ) and normalised absorbance (Y ) for model solutions of desloratadine and their assessment are shown in Table 3. The validation requirements for the linearity of the procedure are met.

Precision and accuracy
The assessment results of the procedure precision and accuracy from the linearity data are shown in Table 4. The validation requirements for precision and accuracy of the procedure are met.

Intermediate precision
Test solutions were prepared by the procedure described in Section 2 Materials and methods. Test portions for the assay on the 1 st and 2 nd days were taken from the same sample of ground tablets. Different analysts conducted an analysis on the 1 st and 2 nd day, using different reagents, volumetric glassware, filters, balances, ultrasonic baths, and spectrophotometers.
The results of the intermediate precision study are shown in Table 5. Pharmacology, Toxicology and Pharmaceutical Science

Robustness -stability of the solutions
For the preparation of the test solution, the ground tablets of desloratadine were used as described in Section 3.

Intermediate precision.
The solutions were stored at 20 °С, protected from direct sunlight.
The results of the stability study of the solutions are shown in Table 6.

Robustness -reliability of the analysis with respect to deliberate variations in the method parameters
The stock solution of desloratadine with concentration about 50 mg/50 mL in 0.1 N·HCl was prepared. The diluents with HCl normality of 0.093 and 0.107 were prepared. The reference solutions for the evaluation of the impact of HCl concentration in the diluent on the reference solution extinction were prepared as follows: 5.0 mL (weighed) of the stock solution were diluted to 250 mL using 0.093 N·HCl, 0.107 N·HCl, and 0.1 N·HCl as diluents. The absorbance of the prepared solutions was measured as described in the analytical procedure. Then, the absorbance of the reference solutions that were prepared using 0.093 N·HCl and 0.107 N·HCl (A init ) were recalculated per the concentration of reference solution with the nominal concentration of HCl (A nom ) by the formula: where m -weights of aliquots of the reference solutions.

Pharmacology, Toxicology and Pharmaceutical Science
The results of the impact of the diluent normality on the desloratadine spectra are shown in Table 7. Three test solutions were prepared using 0.093 N·HCl, 0.1 N·HCl, and 0.107 N·HCl as a diluent. The ratios of absorbance to the test portions were calculated for each test solution (A/m). As was substantiated above, the obtained values for the test solutions prepared using 0.093 N·HCl, and 0.107 N·HCl must not deviate from the value obtained for test solution prepared using 0.1 HCl by more than 0.74 %. The results are shown in Table 8. Evidently, the analytical procedure is robust regarding the deliberate variation of HCl concentration in the diluent.

Discussion
The validation results show that the analytical procedure is suitable for the intended use by the SPhU metrological criteria.
It should be noted that even though the SPhU metrological criteria were developed based on the ICH recommendations, the ICH approach itself does not allow making a conclusion about the suitability of the analytical procedure as it does not specify what the suitability of the analytical procedure for the use for a given application is. Therefore, from our point of view, all validation studies that are based only on ICH recommendations list the research results without the scientifically sound justification of suitability. The concept of uncertainty made it possible to formulate an idea of demonstrating the fitness of the procedure for its intended purpose, based on which the SPhU translated all ICH recommendations into criteria thus providing scientific justification to that idea.
Based on the uncertainty concept, the life cycle approach proposed later by the USP recognised the importance of ICH recommendations but did not translate them into criteria for ICH validation characteristics [17]. As a result, this approach is still separated from the internationally accepted ICH recommendations and even has some contradictions with them.
On the other hand, the USP approach has strongly criticised the one of ICH for being a formal check-box exercise that does not bear in mind the risk of incorrect decision-making in the routine application of the analytical procedure and not leading to the reliable understanding and control of the variability sources [8,17], which, in fact, is also applicable to the SPhU.
In contrast to the ICH approach, USP proposed to use powerful tools for the quality assurance of the analysis result developed for the control of manufacture process (Lifecycle, Target Profile, Risk Management, Control Strategy, Knowledge Management, and Quality by Design) that enormously strengthen the effectiveness of the analytical validation process.
Study limitations. The validation experiment design was developed by the SPhU approach for pharmaceutical preparation «Alerdez», 5 mg desloratadine film-coated tablets produced by «Borshchahivskiy CPP», Ukraine, and might not be suitable for validation of analytical procedures for desloratadine assay in film-coated tablets of other composition and/or manufacture.
The prospects for the further research. Considering those above, further, it seems reasonable to combine the SPhU metrological approach with the USP recommendations for quality assurance of analytical results.

Conclusions
1. The procedure for assay of desloratadine in film-coated tablets «Alerdez» with content limits of ±5 % manufactured by PJSC SIC «Borshchahivskiy CPP», Ukraine, by absorption spectro photometry in the UV region was validated following the recommendations of the State Pharmacopoeia of Ukraine.
2. All validation characteristics and associated performance characteristics were examined in full compliance with ICH recommendations. The analytical procedure has been shown to be suitable for its intended use.
3. The uncertainty of the analytical results and bias ensure acceptable reliability of the decision on medicine compliance with specifications.