1. INTRODUCTION

The presence of impurities may affect the quality, safety, and efficacy of pharmaceutical products. Therefore, the levels in the pharmaceutical products must be controlled within acceptable limits by observing the level of impurities within the permitted daily exposure (PDE), as established by the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) and Guideline for Elemental Impurities (Q3D) [1]. The "Guideline for Elemental Impurities (Q3D)" has been most widely accepted for assessment of metal contamination [2] and, this way, was chosen for metals limits. The EDXRF technique for identifying and quantifying elemental impurities (metals and non-metals) is not indicated as a valid technique by the Brazilian Pharmacopeia (BP) [3], but the United States Pharmacopeia (USP) [4] provides for the use of this technique. According to the BP, the approved techniques for the quantification of heavy metals are limit test and determination by atomic spectrometry (e.g., Inductively Coupled Plasma Atomic-Emission Spectroscopy - ICP-AES or Atomic Absorption Spectroscopy - AAS). The limit test consists of the formation of solid particles of heavy metal sulfides, in suspension, and subsequent visual comparison of the color intensity in the sample and standard preparations in a Nessler tube. The test is semi-quantitative and makes it possible to infer whether or not the sample passes the test, representing the sum of the concentration of contaminants in the sample. The atomic spectrometry method makes it possible to quantify each contaminating element in the sample and different limits are established for each element according to its toxicity, pharmaceutical form, and route of administration. AAS is a quantitative method of metals analysis which was suitable for the determination of 70 elements by three option accessories (flame, furnace, hydride generation) [5] however, many times, needs multiple sample preconcentration processing and expensive instruments [6].

The energy dispersive X-ray fluorescence spectroscopy can measure different elemental impurities, at different concentration levels in various matrices [7]. X-ray fluorescence can be used to quantify practically all elements of interest to pharmaceutical industry, such as heavy metals, within the Maximum Permitted Limits (MPL) as established by the Pharmacopeias assessing the safety of ingestion of each of these elements. More recently, Sauer et al. [8] published a new approach focused on the development of a methodology for screening elemental impurities in solid oral pharmaceutical products using EDXRF methodology with very promising results.

This work is focused on the study of elemental impurities in pharmaceutical products using EDXRF methodology. The elementary impurities used were those belonging to Class 1 (As, Cd, Hg and Pb) and Class 2A (Co, Ni and V), according to the ICH Q3D classification [1]. These two classes (1 and 2) were chosen because are highly toxic to humans. Class 2 impurities are divided into two groups, 2A (Co, Ni, and V) and 2B (Ag, Au, Ir, Os, Pd, Pt, Rh, Ru, Se, and Tl). Group 2A was chosen for this study because having a higher probability of occurrence in the pharmaceutical products. Table 1 shows the MPL of the elemental impurity concentration by the BP [3] and USP <232> [9] for oral intake. The ICH Q3D [1] defines the default values of maximum Permitted Daily Exposure (PDE) for oral pharmaceutical products of Class 1 and Class 2A in units of µg/day, and therefore it is necessary to define the Daily Intake (DI) of the pharmaceutical product to convert to Concentration Limits (CL) in (µg/g). Using Equation 1 below, and considering a daily intake of 10 grams of drug product, is calculated a common permissible target elemental concentration for each component in the drug [1].

This study was conducted considering: Option (1) DI = 10 g/day (maximum daily intake (amount) of the drug product, according to ICH Q3D) and Option (2A) DI = 2.5 g/day (this option is similar to Option 1, except that the drug daily intake is assumed to be less than 10 g) [1]. Considering the value for PDE and DI in Options (1) and (2A), the CL in (µg/g) are displayed in Table 1. In pharmaceutical products, it must be taken into account that the level of impurities cannot exceed the value defined for PDE.

This work brings the validation of the EDXRF method using a matrix based on microcrystalline cellulose and Option 2A for daily intake, focusing on the elements of Class 1 (Cd, Pb, As, Hg) and Class 2A (Co, V, Ni) for elemental impurity control in pharmaceutical products.

2. MATERIALS AND METHODS

2.1. Materials

To perform the analysis via X-ray fluorescence (XRF), pellets were made based on microcrystalline cellulose (brand: Valdequímica, lot nº 023488 and Labsynth, lot nº 227833) spiked with the elements of interest belonging to Class 1 and Class 2A. In some pellets one also used the light elements Na, Mg, Si, Cl, K, Ca, Fe and Ti together with microcrystalline cellulose, in the concentrations needed for the study. The analytes used in this study (light elements and elements of classes 1 and 2A) as reference materials [10] to manufacture the tablets were purchased from Chemplex Industries Inc. (SpectroStandards® XRF reference material preparation kit number 6700). The kit contains 50 different chemical elements, most of which are oxide-based.

In addition to these two groups described above, one also used in some tests a third group of pills purchased in the Brazilian market, containing the following active ingredients: (a) Metformin 500 mg (batch: BR124143, manufacture: 08/2020, expiration date: 07/2023), (b) Simvastatin 20 mg (batch: LD851, manufacture: 01/2021, expiration date: 12/2022) and (c) Glibenclamide 5 mg (batch: ARA06380, manufacture: 11/2020, expiration date: 10/2022). Metformin, Glibenclamide (for the treatment of patients with type 2 diabetes), and Simvastatin (used to treat and prevent hypercholesterolemia) were chosen because they belong to the list of essential drugs and are widely consumed since they are actives for the treatment of diseases that affect a large number of people in the world [11]. Class 1 elements were added because they are the most toxic and thus the most interesting. This third group was spiked with Class 1 elements (Cd, As, Hg, Pb), crushed, homogenized and pressed to form 4 g tablets, similar to the other tablets used in this study. These drugs were purchased in local pharmacies.

2.2. Methods

2.2.1. Pellets preparation

The samples were homogenized using an agate mortar and rotary mixer containing steel balls inside; then, the mixture was separated in portions of 4 g and compacted using a compression machine (pressure of 10 tons and pressing time of 30 s) (brand: Amef, model AP 25 T, electric press), according to the methodology of Marguí et al. [12] with modifications. The final shape of each pellet with 4 g of material (analyte + cellulose) was 30 mm of diameter by 4 mm in height, approximately.

2.2.2. Drug samples

As a way of evaluating how this validation methodology would behave in a real situation, one prepared a recovery test using pharmaceutical tablets produced by Brazilian pharmaceutical companies, which were contaminated with elemental impurities of class 1A (Cd, As, Hg, Pb) only, in high concentration (~35 µg/g) and, in the sequence, the recovery test was performed. The tablets used in this test were (a) Metformin 500 mg, (b) Simvastatin 20 mg, and (c) Glibenclamide 5 mg. The samples were prepared in the same manner as the pellets.

2.2.3. XRF set up

The measurements were carried out using a Malvern Panalytical Epsilon 1, benchtop EDXRF spectrometer, equipped with a 5W, 10kV to 50kV Silver anode X-Ray tube, with energy resolution around 135 eV, having available the following filters for the X-ray beam: Ag, Cu, Ti, and Al. The spectrometer uses a high-resolution SDD (Silicon Drift Detector) detector and operates under atmospheric pressure. The evaluation of elemental concentration of interest elements was performed using the spectrometer configuration displayed in Table 2. In addition to elements belonging to Class 1 and Class 2A, some light elements such as Na, Mg, Ca, K, among others, were also measured in some situations.

2.2.4. Validation

To validate the method for using the EDXRF technique to investigate the elemental impurities of Class 1 and Class 2A, one analyzedselectivity, linearity, precision, accuracy, detection and quantification limits, robustness and interval according to parameters required in the mandatory official compendia [13].

Accuracy

The accuracy was evaluated using certified material with matrix fortification[14]. The accuracy was based on the recovery rate, which is the percentage of the calculated concentration in comparison to the prepared sample concentration (Equation 2), viz.

where R_rec = recovery (%); C_m = mean concentration; C_nom = nominal (true) concentration.

Two groups of samples were prepared. The first group of pellets was prepared using microcrystalline cellulose spiked with elements belonging to class 1 (Cd, Pb, As, Hg) and class 2A (Co, V, Ni), in three different concentrations, viz. i) high (~35 µg/g), ii) intermediate (~15 µg/g) and iii) low (~2 µg/g). The second group of pellets was prepared with the same elements and concentrations, but with the addition of a mix with several light elements (Na, Mg, Si, Cl, K, Ca, Fe and Ti). The light elements were added as a way of modifying the original matrix, creating a "possible" perturbation in the samples. The light elements, that is, elements with a low atomic number are very difficult to measure reliably via EDXRF, but can interfere in the system [15].  So they were added, in order to verify and would be able or not to disturb the analytical system. The light elements added and their respective (high, intermediate and low) concentrations were: Na (0.91; 0.45; 0.072)% (m/m), Mg (0.67; 0.33; 0.053)% (m/m), Si (0.53; 0.26; 0.041)% (m/m), Cl (2.6, 1.3; 0.21)% ( m/m), K (1.4; 0.7; 0.11)% (m/m), Ca (0.82; 0.41; 0.064)% (m/m), Fe (0.8; 0.4; 0.063)% (m/m) and Ti (0.6; 0.3; 0.053)% (m/m).

The accuracy acceptance criteria for elemental recovery according to the working range were 70-150% [7].

Linearity

The linearity was expressed by a linear regression calculated by the method of least squares and by residual analysis. The correlation coefficient (R) is the correlation between the predicted and observed values. This will have a value between 0 and 1; the closer the value is to 1, the better the correlation [16]. The linearity test was conducted using tablets of microcrystalline cellulose spiked with the elements of interest (class 1 and class 2A) at seven concentrations: 70 μg/g, 35 μg/g, 17.5 μg/g, 8 μg/g, 4 μg/g, 2 μg/g and 1 μg/g, and for each concentration the measurements were carried out in triplicate.

Precision

The precision of an analytical procedure is the degree of agreement among individual test results when the procedure is applied repeatedly to multiple samplings of a homogeneous sample. The precision in this study was evaluated by the dispersion of the results, calculating the Relative Standard Deviation (RSD) of the measurement series, as shown in Equation (3), with 9 (nine) determinations being performed, considering the linear range of the analytical method and 3 (three) levels of concentrations: (i) high (~35 µg/g), (ii) intermediate (~15 µg/g) and (iii) low (~2 µg/g), with 3 (three) replicates on each level, for all elemental impurities studied, namely Cd, As, Pb, Hg, Co, V, Ni, viz.

where SD = standard deviation; C_m = mean concentration.

The acceptable relative standard deviation (RSD) for precision parameters considered in this work was: a) concentrations between 1 µg/g to 10 µg/g, RSD < 11%, b) concentrations between 11 µg/g to 100 µg/g, RSD < 7.3% [17].

Detection and Quantification Limits

The detection limit (DL) and quantification limit (QL) are usually calculated using the following set of equations (4):

where DL = detection limit; m = slope of the calibration curve; SD = standard deviation; QL = quantification limit.

However, in quantitative X-ray fluorescence analysis, the DL and QL limits are determined somewhat differently. It is generally accepted that the minimum detectable intensity of a spectral line should exceed by a factor of 3.3 the standard deviation of the integrated background under the spectral line [18]. According to this definition, DL and QL can be expressed in terms of the ratio between the intensity of the background counts (I_b) and the intensity of the spectral line of interest (I₀) (peak counts), multiplied by the initial mass (m_o) used in the analysis, as shown in Equation (5). DL and QL limits were obtained by analyzing a tablet spiked with a mass (m₀) of approximately 2 μg/g of class 1 and class 2A contaminants, in the matrix of cellulose microcrystalline plus light elements. DL and QL depend only on the ratio of background to signal count rates [19, 20].

where DL = detection limit; m_o = mass concentration (μg/g) of class 1 and class 2A contaminants; I_b = intensity of the background line; I_o = intensity of the spectral line of interest (peak counts); QL = quantification limit.

Robustness

The interval must be established from the linearity studies, along with the results from precision and accuracy, depending on the intended application [21]. For determination of impurities, the interval is defined as from the quantification limit up to 120% of the concentration at the specification limit of each individual impurity [22]. In this work, the linear working range was established from the linearity studies, together with the precision and accuracy results for the analyzed values, from 2 µg/g to 35 µg/g.

Statistical analysis

The experimental results were expressed as mean ± standard deviation (SD) or relative standard deviation (RSD), using Microsoft Excel (Microsoft, Redmond WA, USA). The linear regression parameters, fits, statistical calculations such as t and ANOVA tests and graphs shown in this study were performed using Origin Pro 8.5.0 SR1software (www.OriginLab.com).

3. RESULTS AND DISCUSSION

The final goal of the validation of an analytical method is to ensure that every future measurement in the routine analysis will be close enough to the unknown true value for the content of the analyte in the sample [23]. In this way, for evaluation of the methodology the main tests performed were carried out as described before and the results obtained are displayed next.

3.1. Accuracy

The results shown in Tables 3 and 4 indicate that the EDXRF method is very accurate, since the recovery rate for both matrices was within the acceptable range (70-150%) for XRF methods [14].

3.2. Linearity

Linearity should be established across the range of the analytical procedure. The chosen range (70 μg/g, 35 μg/g, 17.5 μg/g, 8 μg/g, 4 μg/g, 2 μg/g) aimed to evaluate linearity in a wide analytical range, because although strict limits of contamination of heavy metals in pharmaceutical products are established, it has been verified that products, particularly vegetable raw materials, exceed the allowed limits [24]. Attention should also be paid to some specific mined excipients, or materials used in very high quantities in a dose form [25].

Figure 1 displays the linear fittings performed between the nominal (true) concentration of impurities and the concentration found (recovered) via EDXRF analysis, in µg/g. The main parameters obtained from the linear regression analyses performed are summarized in Table 5.

The results of the correlation coefficients (R) for all contaminants (Table 5) exhibit a strong linear dependence, since the regression coefficient found for all elements is greater than 0.9963, in compliance with the RDC-166 regulation [13] which states that the correlation coefficient must be above 0.990.

The linear fittings performed (see Figure 1, a,e,i,m,q,u,y) produced correlation coefficients R > 0.990 for all impurities, with this minimum limit being imposed by the RDC-166 norm for the correlation coefficient. According to the results of the statistical analysis of variance (ANOVA) displayed in Table 5, the linear fitting is statistically significant at the significance level of 5%. In addition, the t-test performed allows to conclude that the angular coefficients are ≠ 0 at this level of significance. Moreover, one can easily observe from inspection of Figure 1 that the linear fittings performed were great.

From observation of the data displayed in Figure 1, namely the Standardized Residues versus Found Concentrations (see Figure 1, b,f,j,n,r,v,z), one can clearly see that no possible outliers were found, meaning that none of the data points exhibits a high residue value, with the vast majority of the standardized residues being less than 4. From inspection of the data displayed in Figure 1 (b,f,j,n,r,v,z), one can clearly observe that the data points appear to be randomly distributed, meaning that no clear trends in data distribution such as smile or cone were observed, which is a clear indication that the variance of the experimental errors is homoscedastic. Hence, Figure 1 (b,f,j,n,r,v,z) allows to observe a random distribution in the standardized residue graphs, indicating homoscedasticity, thus confirming the linearity of the method.

Regarding the plots of residues versus collection (sampling) order (see Figure 1, c,g,k,o,s,w,aa),no trends whatsoever could be observed in the data points, meaning that no increasing or decreasing sequences in the data points could be observed and, hence, this is a clear indication that no dependences existed in the observations performed.

In relation to the histograms of the data residues (see Figure 1, d,h,l,p,t,x,ab), a Gaussian distribution centered around zero was expected, without significative deviations to either left or right, which appears to occur with the experimental data presented herein.

The plots exhibited in Figure 1, added to the statistical analysis presented in Table 5, allows to conclude that the experimental errors are independent with normal distribution, and that the variance of the errors is homoscedastic.

3.3. Precision

As shown in Table 6, the dispersion of results for all impurities studied, calculated via RSD, is well within the national norm [13].

3.4. Detection Limit (DL) and Quantification (QL) Limit

The value for maximum permitted limit (MPL) (Table 1) for all impurities of class 1 and class 2A is larger than 2 μg/g. The values obtained for DL and QL (Table 7), indicated that the method can identify and quantify the elements studied in concentrations below the limits stablished in this study and recommended by regulatory agencies. If Option 2A is considered (daily intake of 10 g), the results indicate that only the limit of quantification of cadmium would not be reached.

3.5. Robustness

The robustness of the method was evaluated by measuring the concentration of impurities of elements belonging to class 1 and class 2A, present in the microcrystalline cellulose matrix, acquired from different suppliers and analyzed on different dates. The results obtained are displayed in Table 8.

The results displayed in Table 8 confirm that the method is robust, since it was used cellulose purchased from two different manufacturers and the results obtained for the concentration of impurities are identical for both.

3.6. Interval

In this work, the established linear working range (2 µg/g to 35 µg/g) presented adequate results considering linearity, precision and accuracy.

3.7. Drug samples

The results obtained in the recovery tests can be found in Table 9. Some elements, such as As, presented higher variable recovery, probability due to the analytical process itself, since the X-ray hits on a very small surface. The accuracy of the method was good, however, a deviation of ± 20% can be deemed acceptable [26].

In view of the accepted criteria for the recovery tests (70% up to 150%) [7], the contaminants intentionally placed in the tablet samples were recovered, with the only exception being Simvastatin, with Arsenic (As) slightly exceeding the specification limit for recovery. This result was not expected, since these tablets use in their formulations excipients that are not exactly the same as those used in the test performed and in concentrations that are not known, however, considering the accepted recovery range, one can safely claim that the test results were a success.

4. CONCLUSION

In this research, it was demonstrated that Energy Dispersion X-Ray Fluorescence (EDXRF) technique can be used for detection of elemental impurities of class 1 and class 2A elements in oral tablets. The method proved to be simple, selective, precise, accurate, robust, presenting linear responses for working ranges of interest and low detection and quantification limits (for the group of elements object of this study, LD < 0.4 µg/g and LQ < 1.2 µg/g). It was shown that all validation parameters are satisfied when the daily intake is limited to 2.5 µg/day (Option 1) and remain valid for the maximum daily intake of 10 µg/day (Option 2A), except for the cadmium element, since the limit of quantification for this intake dose should be ≤ 0.5 µg/g and the limit of quantification obtained in this work was 1.2 µg/g.

The EDXRF technique greatly simplifies the chemical analysis of elemental impurities in pharmaceutical products, when compared to the techniques currently accepted by regulatory agencies, such as ICP-MS, AAS, among others, especially in relation to sample preparation.

Acknowledgements

The authors acknowledge CAPES for the Postgraduate Support Program of Community Institutions of Higher Education (Prosuc) in the form of an MSc fellowship granted to Thais Hora Paulino. Project funding by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP, São Paulo, Brazil) (FAPESP Ref. No. 2021/03388-6) is hereby gratefully acknowledged. Funding for Victor M. Balcão through a BPE grant from FAPESP (Ref. No. 2018/05522-9, Project PsaPhageKill) is hereby gratefully acknowledged. This work also received support from CNPq, National Council for Scientific and Technological Development Brazil, in the form of Research Productivity (PQ) fellowships granted to Victor M. Balcão (Refs. No. 306113/2014-7 and 308208/2017-0). The authors have no conflicts of interest whatsoever to declare.

REFERENCES

ICH. International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human use, 2019. Guideline for elemental impurities Q3D. Available at: https://www.ema.europa.eu/en/documents/scientific-guideline/international-conference-harmonisation-technical-requirements-registration-pharmaceuticals-human-use_en-32.pdf., 2019.

KUSHWAHA, P. Metallic impurities in pharmaceuticals: An overview. Current Pharmaceutical Analysis v. 16, n. 8, p 1-9, 2020.

BRAZILIAN PHARMACOPEIA 6th edition Brasília: ANVISA, 2019. Available at: https://www.gov.br/anvisa/pt-br/assuntos/farmacopeia/farmacopeia-brasileira.

UNITED STATES PHARMACOPEIA, 2019. USP <1225> Validation of Compendial Procedures.

ARJOMANDIA, M.; SHIRKHANLOO, H. Review: Analytical methods for heavy metals determination in environment and human samples. Anal Meth  Environ Chem J v. 2, p. 97-126, 2019.

XU, G.; SONG, P.; XIA, L. Examples in the detection of heavy metal ions based on surface-enhanced Raman scattering spectroscopy. Nanophotonics v 10, n.18, p. 4419–4445, 2021.

FURUKAWA, H., ICHIMARU, N., SUZUKI, K., NISHINO, M., BROUGHTON, J., LEINDERS, J., OCHI, H. The comparative verification of calibration curve and background fundamental parameter methods for impurity analysis in drug materials. X-Ray Spectrum v. 46, n. 5, p. 382–387, 2017.

SAUER, B., XIAO, Y., ZOONTJES, M., KROLL, C. Application of X-ray fluorescence spectrometry for screening pharmaceutical products for Elemental Impurities according to ICH guideline Q3D. J Pharm Biomed Anal, v. 179, p. 1–8, 2020.  DOI: 10.1016/j.jpba.2019.113005

UNITED STATES PHARMACOPEIA, 2013. USP <232> Elemental Impurities-Limits.

ISO GUIDE 35:2017, Fourth Edition: Reference materials - Guidance for characterization and assessment of homogeneity and stability Paperback – August 1, 2019.

WHO, WORLD HEALTH ORGANIZATION. Model List of Essential Medicines, 2019. Available at: https://apps.who.int/iris/bitstream/handle/10665/325771/WHO-MVP-EMP-IAU-2019.06-eng.pdf

MARGUÍ, E., FONTAS, C., BUENDI A., HIDALGO, M., QUERALT, I. Determination of metal residues in active pharmaceutical ingredients according to European current legislation by using X-ray fluorescence spectrometry. J Anal At Spectrom v. 24, p. 1253-1257, 2009.

ANVISA. Brazilian Health Regulatory Agency, Resolution of the Collegiate Board (RDC) nº 166, July 24, 2017. Dispõe sobre a validação de métodos analíticos e dá outras providências. Available at: https://www.in.gov.br/materia/-/asset_publisher/Kujrw0TZC2Mb/content/id/19194581/do1-2017-07-25-resolucao-rdc-n-166-de-24-de-julho-de-2017-19194412.

MARSON, B.M., CONCENTINO, V., JUNKERT, A.M., FACHI, M.M., VILHENA, R.O., PONTAROLO, R.  Validation of analytical methods in a pharmaceutical quality system: an overview focused on hplc methods. Quim Nova v. 43, n. 8, p. 1190-1203, 2020.

KONOPKA, J. Options for quantitative analysis of light elements by SEM/EDS. Technical Note 52523. Thermo Fisher Scientific, Madison, WI USA. Available at: https://tools.thermofisher.com/content/sfs/brochures/TN52523_E_0713M_LightElement_H.pdf

BELOUAFA, S., HABTI, F., BENHAR, S., BELAFKIH, B., TAYANE, S., HAMDOUCH, S., BENNAMARA, A., ABOURRICHE, A. Statistical tools and approaches to validate analytical methods: methodology and practical examples. Int J Metrol Qual Eng v. 8, n. 9, p 1-10, 2017. Available at: https://www.metrology-journal.org/articles/ijmqe/pdf/2017/01/ijmqe160046.pdf.

INMETRO. Instituto Nacional de Metrologia, Qualidade e Tecnologia. Orientação sobre validação de métodos analíticos DOQ-CGCRE-008, 2020. Available at: http://www.inmetro.gov.br/Sidoq/Arquivos/Cgcre/DOQ/DOQ-Cgcre-8_08.pdf

VOGEL-MIKUŠ, K., ARČON, I.; KUMP, P., PELICON, P., NEČEMER, M., VAVPETIČ, P., KOREN, S., REGVAR, M. Analytical tools for exploring metal accumulation and tolerance in plants. In: ANJUM, N. A., AHMAD, I., DUARTE, A.C., UMAR, S., KHAN, N.A., PEREIRA, M. E. Phytotechnologies: Remediation of Environmental Contaminants, 2012, p. 468–521. Taylor & Francis, eBooks.

KUMP, P., NEČEMER, M., VEBER, M.  Determination of trace elements in mineral water using total reflection X-ray fluorescence spectrometry after preconcentration with ammonium pyrrolidinedithiocarbamate. X-Ray Spectrom v. 26, n. 4, p. 232–236, 1998.

NEČEMER, M., KUMP, P., VOGEL-MIKUŠ, K. Use x-ray fluorescence-based analytical techniques in phytoremediation. In: GOLUBEV, I.A. (Editors) Handbook of Phytoremediation, 9. ed., 2021, p. 331-358, Nova Science Publishers: New York.

CHOWDARY L. G., RAVISANKAR, P., KUMAR, A. G., MOUNIKA K, BABU P. S. Analytical method validation parameters: An updated review. Int. J Pharm Sci Rev v. 61, n. 2, p. 1-7, 2020.

EMA. European Medicines Agency; ICH Harmonized tripartite guideline - Validation of analytical procedures: text and methodology Q2 (R1), London, 2005. Available at: https://www.ema.europa.eu/en/documents/scientific-guideline/ich-q-2-r1-validation-analytical-procedures-text-methodology-step-5_en.pdf.

GUSTAVO GONZÁLEZ, A., ÁNGELES HERRADOR, M. A. Practical guide to analytical method validation, including measurement uncertainty and accuracy profiles. TrAC - Trend Anal Chem v. 26, n. 3, p. 227–238, 2007.

LUO, L., WANG, B., JIANG, J., FITZGERALD, M., HUANG, Q., YU, Z., LI, H., ZHANG, J., WEI, J., YANG, C., ZHANG, H., DONG, L., CHEN, S.  Heavy metal contaminations in herbal medicines: Determination, comprehensive risk assessments, and solutions. Front Pharmacol, v. 14, p. 1-14, 2021.

DESTEFANO, T., THOMAS, R. From heavy metals testing to the measurement of elemental impurities in pharmaceuticals: Over 100 years in making the change. Spectroscopy  v. 33, n. 5, p. 14-19, 2018.

CROFFIE, M. E.T., WILLIAMS, P.N., FENTON, O., FENELON, A., METZGER, K., DALY, K. Optimizing sample preparation and calibrations in EDXRF for quantitative soil analysis. Agronomy v. 10, n. 1309, p. 1-16, 2020.

Author notes

*jose.oliveira@prof.uniso.br