Post-marketing sampling and testing programs for licensed medicinal products: a narrative review

The globalization of the pharmaceutical market has enabled access to a considerable number of new medicinal products. Consequently, the circulation of substandard medicinal products has also increased. To minimize this problem, post-marketing quality sampling and testing programs are performed to monitor and confirm that the medicinal products available in the market meet appropriate quality requirements. In this review, the post-approval sampling and testing procedures of six regulatory authorities were compared with the goal of strengthening these market surveillance systems. Similarities were observed between the procedures adopted by different regulatory authorities. However, the agencies were not always transparent about the results of these monitoring procedures. A probable mismatch between the registration procedures and the quality requirements listed in official compendiums was observed, which resulted in dissonance and contradiction between the specifications approved by the regulatory authorities and those required in the pharmacopeias. Therefore, strengthening harmonization projects related to these activities can help minimize such difficulties.


INTRODUCTION
Unlike other available items in the market, medicinal products have a very apparent market dichotomy. These are objects of commercial importance and represent one of the sectors that move financial resources around the world, and their reach crosses international boundaries (Lybecker, 2006). It is estimated that the amount of money spent on medicinal products globally will reach more than USD 1.5 trillion by 2023 (IQVIA Institute, 2019). However, medicinal products are public assets that are necessary for the improvement of health policies, and are one of the most tightly regulated products in all countries (Lybecker, 2006).
Since the mid-1930s, many new medicinal products have entered the market. As a result, an outgrowth in the circulation of substandard (authorized medicinal products, but "out of specification") and counterfeit medicinal products (medicinal products deliberately misrepresented) has been observed (Ratanawijitrasin, Wondemagegnehu, 2002;WHO, 2017b). According to the World Health Organization (WHO), it is estimated that about 10% of the medicinal products in the market are substandard (WHO, 2017b). Data from the Promoting the Quality of Medicines (PQM), a United States Agency for International Development (USAID), and the U.S. Pharmacopeial Convention program showed that 5.6% samples of the 15,063 analyzed medicinal products failed the quality test between 2003 and 2013. Depending on the origin of the product, these numbers may be as high as 11.5% (Hajjou et al., 2015). In 2018, an 18.7% prevalence of poor-quality medicinal products was reported in Africa. Although the economic burden of these products is unclear, the estimated impact can range from USD 10-200 billion, depending on the market size (Ozawa et al., 2018).
Treating patients with substandard medicinal products can result in low bioavailability, suboptimal dosing, reduced efficacy, and thus increased mortality and morbidity. It will also lead to resistance to antibiotics, for example, which is one of the greatest threats to public health (Ratanawijitrasin, Wondemagegnehu, 2002;WHO, 2003;Newton, Green, Fernández, 2010;Priyanka, Jayshree, 2016;WHO, 2017a). Depletion of financial reserves of public health institutions, and the loss of confidence in health systems and health care workers are among the other concerning problems (Ratanawijitrasin, Wondemagegnehu, 2002;Newton, Green, Fernández, 2010;WHO, 2017a).
WHO has expressed that countries have to develop regulations and procedures to ensure the safety, efficacy, and quality of medicinal products by controlling and monitoring their circulation in their markets (WHO, 2003). One of these procedures is the sampling and laboratory testing of medicinal products rightfully introduced to the market (WHO, 2016). However, several countries with limited resources have difficulty implementing an effective regulatory system (Ratanawijitrasin, Wondemagegnehu, 2002;WHO, 2003;Rägo, Santoso, 2008). Only about 20% of the countries in the world have well-developed and operational regulation of pharmaceutical products (WHO, 2003;Weyer, 2016).
As a result, regulatory harmonization projects have been an important tool to enhance the quality control programs of marketed medicinal products. (Rägo, Santoso, 2008 (EMA, 2016a).
In addition to the harmonization projects cited above, it is also possible to highlight the regular interactions between regulatory authorities (RAs) through clusters. These are regulatory and scientific information exchanges aimed at leveraging the most successful examples of collaboration, best practices, and to identify opportunities to increase efficiency. This approach addresses a broad range of scientific and regulatory topics, including postmarket actions. Examples are the clusters between the FDA and EMA, which over the years have also had the participation from other RAs (Teixeira, Kweder, Saint-Raymond, 2019).
This review focuses on the procedures used by different RAs to evaluate human medicinal products after marketing authorization, and to identify areas that require considerable attention during harmonization. Consequently, this can contribute towards strengthening the health systems. Therefore, this review is focused on the monitoring of human medicinal products submitted and approved in the marketing authorization applications. Since falsified, unlicensed medicinal products are not covered by the regulations cited in this review, these products must be dealt with separately.

MATERIAL AND METHODS
Five government entities were selected for comparison. Two international reference RAs, the US Food and Drug Administration (FDA) and European Medicines Agency (EMA), a national RA with highly centralized procedures and a regional RA with decentralized procedures (Van Norman, 2016 In addition to these, the WHO guidelines were also analyzed. The post-marketing quality control program guidelines and data were obtained from RAs websites and relevant legislation.
The procedures adopted by different regulatory authorities were compared by analyzing the following: existence of an institutionalized program of sampling and testing medicinal products; similarity in selection criteria for medicinal products to be surveyed; similarity in analytical processes performed during regulatory activity; convergence between approved and official specifications; and transparency of these criteria, procedures, and analytical data.

RESULTS AND DISCUSSION
The regulatory system related to quality assurance of medicinal products involves a wide and varied network of interconnected elements with different levels of complexity. It starts with the technical and scientific evaluation of product development, which is the technical basis for the marketing authorization or licensing. The next phase is the assessment of whether the company's operations are compliant with good manufacturing practices (GMP). The final step is the market surveillance and conduction of surveys to ensure the quality of medicinal products by sampling and testing (EU, 2001).
The analyses performed by the laboratories specified in the general pharmaceutical legislation are essential to ensure compliance with the quality parameters throughout their life-cycle (WHO, 2010).
Therefore, laboratory testing of medicinal products is an integral part of the post-marketing surveillance. It supports the RAs with the inspection services and serves as a basis for subsequent administrative or legal actions (WHO, 2010;FDA, 2015). Thus, the sampling and testing programs involves the selection of products to be evaluated, defining the tests to be performed, the collection of the samples for analysis, and the adoption of legal actions, whenever necessary (WHO, 2016).

Post-market sampling and testing programs for licensed medicinal products
Most of the RAs cited in this review conduct sampling and testing of licensed medicinal products through periodic programs, such as "Sampling and Testing Programme" conducted by EMA (EMA, 2016c), the "market supervision program" executed by INFARMED (Infarmed, 2018), the "DeMuestra La Calidad" program of INVIMA (Invima, 2017), and the "Drug Quality and Sampling Testing Compliance Program" of FDA (FDA, 2015). In contrast, the Brazilian program is non-periodic, but performs routine market supervision (Brazil, 1973;Brazil, 2013).
Nonetheless, the regulations enforced by the FDA and EMA differ. While the FDA has a strictly centralized process, the EMA is decentralized through a network of nationally centralized agencies with synchronized regulations preserving national "autonomy"; post-market testing of medicinal products is not an exception (Van Norman, 2016). In the American market, medicinal products are tested in FDA laboratories or its designated laboratory (FDA, 2018). EMA uses the expertise and laboratories under the responsibility of EU member countries, the Official Medicines Control Laboratories (OMCL) network. All member states mutually recognize the laboratory results without further confirmatory testing (EDQM, 2016a).
A regulatory system similar to FDA is observed in INVIMA and INFARMED, when Portuguese RA is observed off the topic of the EU. Both are national RAs accountable to evaluate, authorize, control, and test the medicinal products introduced in the market, also acting as Reference Laboratory on the Quality Control of Medicines (Portugal, 2012;Colombia, 2015).
However, according to the Colombian procedures, when a non-compliance event is reported, the manufacturer's laboratory is responsible for conducting a retention sample testing under the supervision of INVIMA. Colombian RA maintains that such procedure is necessary to ensure if the identified non-compliance is due to the manufacturing process or due to transportation, storage, method of analysis or other correlated factors (Invima, 2017). Edvaldo Oliveira Neves, Paloma Michelle de Sales, Dâmaris Silveira In these analytical aspects, the Brazilian RA and EMA have characteristics in common; since ANVISA does not have a laboratory it relies on the expertise of laboratories from other levels of the Federation or seeks assistance from a specialized laboratory (Brazil, 2015;Brazil, 2017).
Owning a laboratory is not a requirement for carrying out post-marketing sampling and testing programs. It is possible to hire external laboratories. However, it can be a limiting factor. Some common barriers to implement these programs in countries with limited resources are poor laboratory infrastructure, lack of equipment, untrained personnel, difficulties in acquiring the reagents and supplies required to perform quality control analysis (WHO, 2010;Pribluda et al., 2014).
Thus, sharing of analytical data can minimize this difficulty. Though this kind of sharing is on the agenda of the harmonization projects, it is still far from being realized, except for EU. Perhaps, it may be due to the fear of loss of autonomy by many countries, limitations in the integration of national and regional databases, or even difficulties in recognizing the identity of medicinal products since RAs cannot always be sure they are referring to the same product (EMA, 2016a). Table I shows test results from the sampling and testing programs mentioned in this review. It is important to note that despite the current discussion on transparency and access to information, there is limited access to data related to quality of medicinal products. Although many technologies could be used to improve transparency, it has been found that they are not used efficiently. Data are not up to date and do not provide sufficient information for generating reports or surveys. This includes laboratory data, guidelines, and procedures. Thus, there is inefficient communication of RAs with the society.

Selection of medicinal products or groups of products and manufacturers
Products included in the post-marketing quality control testing programs for medicinal products are selected through a risk-based approach. Testing focuses on medicinal products, holders of the marketing authorization or medicinal product manufacturing companies, dosage forms, and active pharmaceutical ingredients (API), for example, using targeting products that are most likely to pose a risk to patients. Examples include products indicated by the offices based on their regulatory intelligence: narrow therapeutic index drugs, products whose consumers reported complaints, companies with a past history of non-compliance, produced by manufacturers for whom there is sparse evidence of compliance with the principles of GMP (EDQM, 2014;FDA, 2015;Anvisa, 2016;Sindusfarma, 2016;Invima, 2017;Infarmed, 2018;WHO, 2018a).
It is important to note that the manufacturers and the selection of their products based on respective analytical history can be an important tool for the risk-based approach. It has been identified that some manufacturers tend to produce drugs with API values towards lower tolerance limits. Although still within tolerance limits, this indicates deficiencies in compliance with GMP (Neves, de Sales, Silveira, 2019).
This risk-based approach is reinforced by WHO guidelines, which emphasize that laboratory testing should focus on medicinal products or companies most likely to pose risk to patients, and it should be performed if it adds value to the evaluation since these analyses demand considerable resources (WHO, 2018a). Moreover, this risk-based approach allows the laboratories to concentrate their limited resources on those medicinal products considered to have a high risk of quality defects (EDQM, 2016b). The use of this approach is a common point among the cited RAs.

Quality specifications in post-market quality control
As mentioned earlier, laboratory analyses should add value to the evaluations performed. Therefore, the specifications to be applied must be in line with the answers sought (WHO, 2018a). Specification, by definition, contains a list of tests, analytical procedures, tolerances limits or acceptance criteria to which product should conform (ICH, 1999).
Thus, evaluating impurities or dissolution without performing the assay is not reasonable. Likewise, if the purpose is to verify the compliance of a product with registered specifications, the information approved by RAs or regulatory specifications should be used. However, if products containing the same API and dosage from different manufacturers are to be evaluated, pharmacopeial or official specifications are recommended (WHO, 2018a).
Considering this, although the pharmacopeial specifications are a public statement of the appropriate quality adopted in a country or region, i.e., their regulatory expectations, they are not always considered a conclusive quality requirement. The analysis performed using the pharmacopeial specifications should be seen as quality screening (Calam, 1995;Conceição et al., 2014;EDQM, 2014;WHO, 2016;Heyward et al., 2018). However, by applying a strong official quality standard, nonconformity in a public monograph can be a fast and efficient way to remove a substandard product from the marketplace, as opposed to lengthy administrative or legal action (Heyward et al., 2018).
The specifications developed by the manufacturers involve several factors, such as dosage form, clinical aspects of the API of medicinal products, critical attributes, parameters of the product related to manufacturing process, statistical sampling plans, among others (Dong, Tsong, Shen, 2015;Burdick et al., 2017). In contrast, pharmacopeial specifications are supported in literature reviews and marketing authorization data. In light of this, official quality standards are likely to have broader specifications than those adopted by manufacturers (FDA, 1987;EU, 2001;Portugal, 2006;Anvisa, 2013 For marketing authorization or licensing, the FDA requires the manufacturer or marketing authorization holder to present the relevant quality requirements throughout a product's shelf life. Broader limits than those in the official specifications are not ordinarily approved as regulatory specifications (FDA, 1987;Wechsler, 2002). Additionally, some guidelines promote convergence between regulatory specifications and monographs in the United States Pharmacopeia (USP), to minimize discrepancies between them (FDA, 2014;USP, 2016;USP, 2017;USP, 2018). Thus, FDA considers USP specifications in its medicinal products sampling and testing program. If a particular monograph is not specified in the pharmacopeia or in the occurrence of outof-specification (OOS) results, regulatory specifications are performed (FDA, 2015;FDA, 2018). Figure 1 shows the general process of medicinal products sampling and testing as well as highlights the need for convergence between regulatory and official specifications. GMP, Good manufacturing practices; RAs, regulatory authorities; OOS, out-of-specification.
Post-marketing sampling and testing programs for licensed medicinal products: a narrative review European RAs expect more restrictive in-house product release specifications than regulatory and official specifications. Many manufacturers implement in-house fairly tight limits that provide more assurance that their product will remain within established acceptance criteria throughout its shelf life (Wechsler, 2002). Therefore, in both scenario European or American, pharmacopeias describe the minimum and mandatory requirements applicable to all medicinal products placed in the market. However, a more restrictive analytical profile is discernible in the European market.
Nevertheless, European RAs post-marketing sampling and testing programs do not always adopt pharmacopeial specifications as the screening tool. INFARMED, for example, performs direct regulatory specifications as a starting point in the market surveillance national program (Infarmed, 2018). As a form of screening, EMA sampling and testing program evaluates EMA centrally authorized products, which is carried out based on its own analytical protocol, the "Common Test Procedure." This protocol is established based on a defined list of parameters to be tested including the variety of regulatory specifications, API, variations in the dosage form, specifications from European Pharmacopeia, and OMCL in-house methods. With OOS results, new analyses are performed using the regulatory specifications for the sample in question (EDQM, 2014).
It is important to mention that the European Pharmacopeia has a limited number of specific monographs for finished products. Among the almost 2400 existing monographs, only 1% is related to finished products. It is a compendium that only describes the general requirements for the different dosage forms and specific monographs of API (EDQM, 2017;Conceição et al., 2019). This justifies disuse of compendium as a primary reference for the analysis of finished products.
The Brazilian and Colombian RAs, on the other hand, require manufacturers to meet specifications of the recognized pharmacopeias in the Country (Colombia, 1995;Anvisa, 2009;Invima, 2017). It is important to mention that besides Brazilian Pharmacopeia, several other Pharmacopeias are recognized in Brazil: German, US, Argentine, British, European, French, International, Japanese, Mexican, and Portuguese (Anvisa, 2009). Colombia recognizes US, British, French, German, European, and International Pharmacopeia (Colombia, 1995). However, this country does not have its official compendium (Colombia, 1995;WHO, 2018b).
The Brazilian guidelines related to monitoring the quality of medicinal products is scattered and sometimes with divergent outlooks. In this manner, these guidelines are not clear about the definition of quality specifications and their applications in this conformity assessment process (Brazil, 1976;Anvisa, 2009;Anvisa, 2010). Also, there are no procedures that converge the approved specifications with that described in the national pharmacopeia (Lima, 2017;Neves, de Sales, Silveira, 2019). Thus, although this convergence seems to be well established in the American regulatory system, it is not institutionally planned in the Latin American context (FDA, 2014;USP, 2016;USP, 2017;USP, 2018). Non-harmonized quality testing standards can lead to substandard products remaining in the market with questionable clinical efficacy and, not infrequently, judicialization of regulatory actions (Heyward et al., 2018).
In this context of diversity of pharmacopeias, it is observed that there is a certain similarity between the Latin American and European scenario. Although there are instruments for the mutual recognition of analytical data and product licensing in the EU, to our knowledge, there is no available tools to promote convergence between analytical specifications in the individual monographs of the 16 national pharmacopeias in the European block (EDQM, 2016a;EMA, 2016b;WHO, 2018b). Considering this, if the marketing authorization holder chose to market the product in different countries of the EU, it will have to observe different quality metrics established in these different compendiums (Bouin, Wierer, 2014).
From the perspective of the RAs, it is not reasonable to authorize a product or supervise the market using different rules for the same product. Therefore, it seems logical to establish guidelines for the use of those compendia when the same product is described in more than one pharmacopeia, which are recognized in the country or region and serve as the basis for the registration of a product whose circulation will be beyond national borders. Table II compares the profile of post-authorization sampling and testing programs considering the information already described. Although significant similarities are seen in the selection criteria of the samples, there are some major differences in the screening process at the analytical stage. It is important to note that in the context of the diversity of pharmacopeias, clear criteria have not been found (and it is unlikely that they exist) when more than one official compendium exist. Despite this, all RAs perform regulatory specification for final decisionmaking in the case of OOS results.    -(Brazil, 1973;Brazil, 1976;Anvisa, 2009;Anvisa, 2010;Brazil, 2013;Brazil, 2015;Anvisa, 2016;Sindusfarma, 2016;Lima, 2017;WHO, 2018b;Anvisa, 2019;Neves, de Sales, Silveira, 2019) and General guideline of the National Program for Quality Control of Medicines -PROVEME unpublished work; 2 - (EU, 2001;Wechsler, 2002;EDQM, 2014;EDQM, 2016a;EMA, 2016c;EMA, 2016b;WHO, 2018b); 3 - (EU, 2001;Portugal, 2012;Infarmed, 2017;Infarmed, 2018;WHO, 2018b); 4 - (Colombia, 1995;Colombia, 2015;Invima, 2017;WHO, 2018b); 5 - (FDA, 1987;Wechsler, 2002;FDA, 2014;FDA, 2015;USP, 2016;USP, 2017;FDA, 2018;USP, 2018;WHO, 2018b).

Other regulatory tools
In addition to the sampling and testing programs described here, other tools can be used as auxiliaries for market surveillance. One of these tools is the "Three-Level Approach", proposed by the PQM program. This approach uses three interlinked approaches with increasing level of complexity, to strengthen the quality assurance systems of medicinal products. The first level includes a visual and physical inspection to assess the packaging conditions and the physical characteristics of the actual medicinal products. The second level consists of rapid analytical tests that evaluate a limited number of quality attributes and can be easily performed by trained personnel. The third level involves quality control testing by the regulatory specifications and is performed at an appropriate laboratory. This approach is applicable mainly to countries with limited resources, laboratories with poor infrastructure, and those with difficulty in structuring a robust medicinal products monitoring program in their market (Pribluda et al., 2014).
Example of other tools are portable devices to track substandard medicinal products. There are a multitude of devices for this purpose. However, these tools lack costeffectiveness, and accuracy and specificity of scientific data. Moreover, some devices require processing of complex and specific calibration models. These factors limit the use of this tool (Vickers et al., 2018).

Limitations
This study was supported by public information from RAs. In doing so, it was difficult to obtain clear information on these procedures. Thus, some of the regulatory guidelines may not have been identified during this review. Furthermore, there are regulatory procedures that are not publicized or interpretations not explicit in the guidelines, which could not be validated since they cannot be referenced. Therefore, although regulatory transparency should be an essential feature of RAs, this requirement remains fragile and fragmented. This hampers access of information to all stakeholders, including regulators, the regulated sector, health care professionals, academics, or the community (Sousa, Ramalho, Silveira, 2016).

CONCLUSION
It has been observed that the procedures for sampling and testing of authorized medicinal products are very similar among RAs, although limitations in transparency regarding the implementation of these procedures in some countries are evident.
For countries or regions where more than one official compendium is recognized, there is a need for greater clarity in the procedures used for market surveillance and guidelines for the use of these pharmacopeias. It is also worth emphasizing the need to frequently update the national compendium, when it exists, to avoid discrepancies between the specifications approved by the RAs and those that are officially required.
Therefore, it is important to strengthen the actions developed by EDQM in the EU and PANDRH in Latin America. Moreover, the actions related to the Mercosur Pharmacopeia, that was created as a tool for regulatory strengthening and pharmaceutical production in South America, but which has been stalling for a long time are also factor that can contribute to the improvement in regulatory standards in Latin America.
In conclusion, sharing analytical data in Latin American region to strengthen the monitoring of the quality of medicinal products circulating in the region or as a tool to reduce the excessively high costs, is inherent to these analyses.