Elemental Impurity Analysis

The United States Pharmacopeia (USP) has revised its elemental impurities limits and procedureal chapters with implementation set for May 2014. The author explains the need for these revisions and provides a look at some of USP's proposed techniques for elemental impurity detection and identification.

Revisions to US Pharmacopeia General Chapter <231> Heavy Metals have been mooted and proposed for more than a decade, and it has long been known that the current methods are highly subjective and likely to prove inaccurate, at least for certain metals. The road to reform has been somewhat stuttering, but after a long period of review and commentary, on Dec. 1, 2012, General Chapter <232> Elemental Impurities—Limits and Chapter <233> Elemental Impurities—Procedures will be published in the second supplement of the US Pharmacopeia 35–National Forumulary 30 (USP–NF).

On May 1, 2014, when USP 37–NF 32 becomes official, all references to Chapter <231> will cease to exist, and conformance to Chapter <232> and Chapter <233> within the General Notices will be required. The acceptance of these chapters will open the door for laboratories to use a wider range of methods for analyzing heavy metal contaminants. Of course, these methods will still need to be validated, and there may still be room for debate about which methods are best for any given situation, but at least the dubious methods of Chapter <231> will cease to be available for medicines marketed in the United States. (For the time being, the comparable methods used in the European and Japanese pharmacopoeias will continue to be available.) This paper addresses the need for new compendial requirements, with a focus on elemental impurity detection and identification.

The need for change

Testing for heavy metals is actually one of the most established ideas contained within the national pharmacopeias around the world. In fact, USP has included a general test for heavy metals since 1905 in the eighth volume of the pharmacopeia, which used sulphide precipitation to detect antimony, arsenic, cadmium, copper, iron, lead, and zinc. As it happens, the purpose of the test had more to do with prevention of mislabeling than prevention of contamination, because heavy metal salts were often used in therapy and one had to know which salts were present in a treatment. The need to detect residual contamination was established in 1942, with the introduction of USP volume XII, in which a lead-containing standard was included in the test. The goal was to detect potentially poisonous heavy metal residuals, such as lead and copper, because these metals were widely used in production equipment at the time. Interestingly, metals such as iron, chromium, and nickel were not revealed by the test (1). Ultimately, it is the inapplicability of a "standard" test (such as that defined by Chapter <231>) that has led to its demise and the need for more flexibility.

Industry knowledge of common metal contaminants. Metal impurities are rightly a cause for concern in pharmaceutical products and there are many means by which a product might become contaminated. There are many inorganic impurities that are deliberately added to the pharmaceutical processes (e.g., catalysts). There are other impurities that can arise as undetected contaminants from starting materials or reagents, or that come from the process itself (e.g., leaching from pipes and other equipment). Then, of course, there are metal ions that occur naturally within the plant or mineral sources that are used to produce the active ingredients of pharmaceuticals and herbal medicines.

Regardless of how metals may get into a product, or previous certification of these metals, pharmaceutical producers must carry out tests to demonstrate the absence of impurities before using materials in a pharmaceutical product.

General Chapter <232>: New limits

USP General Chapter <232> Elemental Impurities—Limits sets out the acceptable levels of 15 elements in final drug products. These limits have been evaluated from toxicological data and are expressed in terms of a daily permissible exposure (DPE) limit. The DPE also takes into consideration the route of administration (e.g., oral, parenteral, or inhalable) with orally administered drugs having a higher permissible limit than parenteral or inhaled drug products. Where elements on the list are known to be present or have the potential to be present then compliance with the specifications must be assessed. The 15 elements addressed in Chapter <232> are based on the International Conference on Harmonization's (ICH) Q3D Elemental Impurities Working Group pre-Stage 2 draft guideline (2).

Chapter <232> covers arsenic, cadmium, mercury, and lead— all elements that are considered ubiquitous and therefore must be assessed in all cases. In addition, the chapter covers iridium, osmium, palladium, platinum, rhodium, ruthenium, chromium, molybdenum, and nickel. The second group of elements may be present in products as a result of being added deliberately, for instance, in the form of a catalyst or through interactions with metal components through the manufacturing process.

Because the ICH Q3D guideline is still being reviewed and is likely to expand to cover more elements, it has been decided that a review of Chapter <232> will happen after the deliberations on ICH Q3D guidelines have been completed. At this stage, the scope of Chapter <232> may be expanded to cover more elements, or an informational chapter may be incorporated to cover elements of low toxicity.

The outdating of long-standing tests

It is fair to point out that the methods of USP General Chapter <231> were developed before the introduction of modern analytical instruments. These methods were easily transferable from one laboratory to another and did not require sophisticated instrumentation or specialized expertise. Hence, a competent laboratory staff member could perform the same techniques with relative ease. The problem was that the methods themselves were flawed, no matter how competent the analyst.

For example, Chapter <231> methods involved subjective visual examination and comparison of the sample solution with a lead standard. Similar to the method of 1905, the compendial methods used a reaction to form the sulphide of any metal ions present and the total metal content was reported against the lead standard response as a limit test.

The validity of this comparison relied on several assumptions, all of which can be questioned. For example, the compendial method assumed that each of the heavy metals in the sample matrix would react in a like manner to lead to form a sulphide species. This assumption applied despite many sulphides being known to be insoluble and despite some elements being known to have a far more intensely colored sulphide than the lead standard against which it was being assessed. Similarly, the compendial method assumed that the reaction kinetics for lead sulphide would be very similar to that of the other metal sulphides and that reaction kinetics were not greatly affected by the sample matrix. A final major and unsafe assumption was that the heating and/or ashing step of the method would have no impact on volatile metals (3).

Work has been carried out that suggests that recovery of mercury can be as little as 2% using the <231> compendial method, which clearly introduces a massive error in the final result (2). Other laboratories have reported similar poor recovery of metals such as tin, selenium, and antimony. These examples are by no means the only reasons to challenge the validity, applicability, and reliability of the compendial methods. In fact, additional chapters for the control of specific metals and other inorganic impurities have been added to USP over the years. Significant among these additions has been USP Chapter <730> Plasma Spectrochemistry, which gave laboratories the opportunity to use techniques such as inductively coupled plasma with either mass spectrometry or atomic emission spectroscopy (ICP–MS and ICP–AES).

The advantage of ICP methods is that they can provide specific detection and quantification for each of the elements specified in Chapter <232>. The subjectivity of the semiquantitative comparison that is required by the compendial methods is eliminated with ICP. The ICP techniques are also quicker in most cases, requiring a smaller sample size and giving a better detection limit for all the elements of interest. The sample preparation method for ICP, for example, is less likely to lead to the loss of the volatile elements.

Chapter <233>: New techniques

USP General Chapter <233> Elemental Impurities—Procedures sets out the general conditions for testing, covering preparation, analysis, and the parameters for validation. The preparation methods referred to above are neat, direct aqueous solution, direct organic solution and indirect solution.

Neat samples are in such a state that they can be used without further preparation. More commonly used solutions will need to be prepared prior to analysis, and the simplest of these procedures is preparation of a direct solution whereby a product is dissolved or diluted with water/dilute acid or an organic solvent to give a solution for analysis.

In many cases, it may be desirable to treat the sample by breaking down any organic material contained within it; such a step typically reduces the ffect of the matrix effect which might otherwise give rise to false positive/negative results. If a sample is prepared in this way, then it is referred to as an indirect solution. These solutions are generally prepared using a microwave digester. In this technique, a small amount of sample is weighed into a vessel and acid is added. The vessel is sealed and placed into a microwave. In the microwave, the sample is heated to temperatures of up to 250 °C and pressures of up to 55 bar. Under these conditions, the sample matrix is effectively destroyed and the metal atoms are released into solution. After the sample is cooled, it is made up to a suitable volume with water ready for analysis.

ICP–MS and ICP–AES. As noted above, Chapter <233> sets out two procedures for analysis, ICP–MS and ICP–AES. The latter is also sometimes referred to as ICP–OES, which stands for optical emission spectrometry. In this technique, the sample solution is fed into an argon plasma which has a temperature of approximately 10,000 °C. The sample matrix is destroyed under these conditions, and individual atoms are released. These atoms are then excited to a higher energy state. As the excited atoms cool, they return to a "ground state." The process releases energy in the form of light, the wavelength of which is specific to a particular element. When this light falls on a detector, it can be quantitated and the amount of analyte can be evaluated.

ICP–MS is the second procedure specified in Chapter <223>. This technique also uses a plasma, but with this technique, the plasma is used to ionize the metal atoms which are then fed into a quadrapole which separates the ions according to their mass-to-charge ratio. Following separation, the ions fall onto a detector and the sample can be quantified.

Differentiating the new techniques. Both ICP–AES and ICP–MS are able to analyze several elements simultaneously. As a result, sample throughput can be very quick, typically 2–3 minutes per sample. Generally, it is fair to say that ICP–AES instrumentation is cheaper than ICP–MS, but both instruments have relatively high running costs due to the consumption of argon in the plasma. The key difference between the instruments is the detection limit. The ICP–MS typically has detection limits 100–10,000 times lower than that of ICP–AES. Both techniques are capable of analyzing to the levels required by USP, but ICP–MS can offer a much lower detection limit. Chapter <233> states that for both techniques, steps can be taken to remove matrix interferences. For ICP–AES, these interferences can occur from overlapping wavelengths. In this case, alternative wavelengths can be used for analysis. Also, many instrument manufacturers have correction techniques built into the operating software.

In the case of ICP–MS, the sources of matrix interferences come from the fact that different species can have the same mass/charge ratio. For example, argon chloride appears at the same mass as arsenic, giving false positive results. To remove these interferences, many instrument manufactures use special cells within the instrument that can add gases to the ions and mitigate the interferences.

Alternative methods

Other techniques can be used in the analysis of elemental impurities, but each must be validated to ensure that it is suitable and able to detect the analytes at the required level. Below are a few options:

Flame atomic absorption spectrometry (FAAS): This simple and relatively cheap technique has relatively high detection limits, especially for elements such as mercury and arsenic. FAAS can only analyze one element at a time.

Vapor generation atomic absorption spectrometry (VG–AAS): This technique involves a chemical reaction to release metals in the form of gaseous hydrides. It has improved detection limits compared with FAAS but can only be used for arsenic, bismuth, germanium, lead, antimony, selenium, tin, and tellurium. Only one element can be analyzed at a time. Reagents are used to generate the hydride therefore generating a higher cost than traditional AAS.

Graphite furnace atomic absorption spectrometry (GFAAS): In this technique, a small amount of sample is slowly heated to first dry then ash the sample. Thereafter, the temperature is raised very rapidly to volatilize the metal of interest. This technique has very good sensitivity and can be used to look at very low levels of analyte similar to those achieved by ICP–AES, but is prone to chemical interferences affecting the results. Also the analysis is slow and can be costly.

The demise of Chapter <231> means that modern techniques referred to above, and others, will now come become more common, and the old wet chemistry results will cease to be valid.

Conclusion

One can only sympathize with the scientists at USP that have responsibility for the standard pharmacopoeial methods involving heavy metals. Of the 4000-plus monographs in the USP–NF, there are approximately 1000 that specify a limit of heavy metals, in either a drug substance, excipient, or drug product (4). December 2012 marks the beginning of the end for Chapter <231> and the introduction of Chapters <232> and <233>. By May 2014, <231> will cease to exist, and by this point, validated procedures need to be in place to cover the removal of Chapter <231>. The 18 months between these dates may seem like a long time, but considering the number of existing monographs that contain the limit of heavy metals test, this timeframe seems very short. Overall, the USP changes, although daunting, can lead to improvements for the industry, including by better protecting the public through effectively tested medicines. Manufacturers will have peace of mind that they are providing clean and safe products to the market.

Alan Cross is a scientist at RSSL, Reading Science Centre, Whiteknights Campus, Pepper Lane, Reading, Berks, RG6 6LA, UK, tel. +44 (0)118 918 4129, enquiries@rssl.com.

References

1. O. Pedersoen, Pharmaceutical Chemical Analysis: Methods for Identification and Limit Tests (Taylor & Francis, 2006).

2. N. Lewen et al., J. of Pharm. and Biomedical Anal. 35 (4) 739–752 (2004).

3. ICH, Q3D Impurities: Guideline for Metal Impurities, Final Concept Paper (2009).

4. D.R. Abernethy, Chief Science Officer, USP, presentation online at www.usp.org.