The long awaited revision

A pharmaceutical product can become contaminated by metal impurities via a number of ways. There are many inorganic impurities that are deliberately added to the process (e.g., catalysts), while undetected contaminants from starting materials or reagents, or that come from the process, such as leaching from pipes, mixing vessels and other equipment, can raise serious issues. There are also naturally occurring metal ions within plant or mineral sources that are used to produce APIs and herbal medicines.

Peter Rooney

Regardless of how they might get into a product, the control of these impurities may be certified by a vendor, but pharmaceutical producers must conduct their own tests to demonstrate the absence of impurities before using these materials in a manufactured article.

Long established requirement

Heavy metal testing is by no means a recent phenomenon. The European Pharmacopoeia (Ph. Eur.) 2.4.8 was derived from the heavy metal tests inherited from national pharmacopoeias, many of which had been in existence for years. The US Pharmacopeia (USP) has included a general test for heavy metals since Volume VIII of 1905, which used sulfide precipitation to detect antimony, arsenic, cadmium, copper, iron, lead and zinc. The purpose of the test had more to do with prevention of mislabelling than prevention of contamination as heavy metal salts were often used in therapy, and which salts were present in a treatment had to be known. The detection of residual contamination was introduced in 1942 with Volume XII, in which a lead-containing standard was included in the test. The aim was to detect potentially toxic heavy metal residuals, such as lead and copper, because these were widely used in production equipment at the time. Interestingly, metals such as iron, chromium and nickel were not revealed by the test.

Limitations

It is the limitation of the current 'wet chemistry' methods described in USP Chapter <231> that has led to the decision to revise the chapter.

Current compendial methods were all developed before the introduction of modern analytical instruments. Their advantage is that the methods are easily transferable from one laboratory to another and do not require sophisticated instrumentation or specialist expertise. However, the analysis also involves subjective visual examination and comparison of the sample solution with a lead standard. As with the method of 1905, the compendial methods use a reaction to form the sulfide of any 'so called heavy' metal ions present and the total metal content is reported against the lead standard response as a limit test.

The validity of this comparison relies on several assumptions, all of which can be questioned. For example, the compendial method assumes that each of the heavy metals in the sample matrix will react in a like manner to lead to form a sulfide species. This assumption applies despite many sulfides being insoluble and some elements having a far more intensely coloured sulfide than lead. Similarly, the compendial method assumes that the reaction kinetics for lead sulfide will be very similar to those for other metal sulfides and that reaction kinetics are not greatly affected by the sample matrix. A final major assumption is that the heating and/or ashing step of the method will have no impact on volatile metals or metal compounds.

These are not the only reasons to challenge the validity, applicability and reliability of the compendial methods. It is no surprise that additional chapters for the control of specific metals and other inorganic impurities have been added to the pharmacopoeia during the years. Significant amongst these in the USP is Chapter <730> Plasma Spectrochemistry, which was introduced approximately 2 years ago and gives laboratories the opportunity to use techniques such as inductively coupled plasma with either mass spectrometry (ICP-MS) or atomic emission spectroscopy (ICP-AES).

The advantage of ICP methods is that they provide specific detection and quantification for each of the elements expected to give rise to a positive response in the compendial methods, such as arsenic, selenium, lead, silver and bismuth. The subjectivity of the semiquantitative comparison that is required by the compendial methods is eliminated with ICP. In some cases, ICP is also quicker, requires a smaller sample size and depending on sample preparation may give a higher recovery of some or all of the elements of interest. Additionally, the sample preparation methods that can be used for ICP is less likely to lead to the loss of the volatile elements.

Proposed revisions

The proposed revision to Chapter <231> recommends procedures that rely on modern analytical technology and includes limits that are based on toxicity and exposure levels for the selected metals. In its stimuli article for the proposed revision, the USP Ad Hoc Advisory Panel on Inorganic Impurities and Heavy Metals and the Expert Committee proposed that "the selection of an instrumental technique and a procedure for the evaluation of the inorganic impurities specified ... requires the evaluation of a large number of variables including, among others, sensitivity, precision, accuracy, compatibility, time, and cost. The method selected may include plasma spectrochemistry, atomic absorption spectroscopy, or any other method that displays requisite accuracy (trueness and uncertainty) and established sensitivity and specificity."¹

In enabling heavy metal limit testing to be conducted using ICP-MS and other plasma spectrochemistry, the revised USP <231> will not entirely rid itself of problems as the stimuli article raised several questions, which any revision will need to address.

Room for improvement

Any standard method always runs the risk of being too general and unable to deal with particular circumstances. The proposed revisions to Chapter <231>, as outlined in the stimuli article, do raise some concerns; for example, certain elements have not been included in its Table 1, which sets out the proposed element limits for oral and parenteral materials.

Silver is not in the list of elements in the table, which is surprising given that silver would have been detected by the methods given in Chapter <231>. With the current changes to water purification processes, many industries use sacrificial copper/silver anodes to kill microbial contamination and so one might have expected silver to be included in a list of elements having toxicological properties.

Other elements have also been omitted from Table 1, such as titanium, barium and zirconium, that have toxic forms. It is feasible that the USP intends that the individual limit tests for barium will remain as they are now, rather than being incorporated into this revision. However, given that rarely used elements such as iridium and thallium have been included, it is curious that other elements such as those listed above, as well as germanium and gallium, both of which are toxic, have not been included.

From a practical perspective, there is also room for debate regarding the reliability of the proposed methods; for example, it would appear that spiking of samples is being proposed as a means of confirming recovery. However, where the concentration of certain metals is already high, the spiked amount may be too small an increase to be detected above the level already found in the sample. In extreme cases, the contribution of the spike could be within the tolerances of normal result variability. If, for example, the method prescribed in the stimuli article were to be applied to ferrous fumarate to determine the levels of mercury, arsenic and lead, the system suitability would not be achievable for iron. This observation means there will have to be specific exclusions in the chapter or individual monographs to cope with such occurrences.

The stimuli article illustrates a flow chart for determining the sample preparation method. However, no acknowledgement is made that with a water soluble sample, solution viscosity and salt strength must also be taken into account, as both of these properties will have severe effects on the performance and life of the nebulizer in the ICP instrumentation.

In the case of water insoluble samples, decisions must also be made about which solvents are appropriate to use. For the time being, the stimuli article makes no recommendations and proposes no list of acceptable/unacceptable solvents. Again, this choice will have major impacts on the cost of the analysis and lifetime of equipment. Should the method force the use of analysis in an organic solvent, it may oblige testing laboratories to buy solvent system nebulizers and plasma torches as standard torches are often unsuitable where organic solvents are involved. There is also an assumption that the elements of interest will be soluble in the solvent of choice. This may lead to instances where known sample solubility data are used to select a solvent, but the trace metals will not be solubilized and because of the low concentrations this will not be apparent as out of solution. Interestingly, the option of microwave digestion is only offered when a sample is deemed to be insoluble in water or 'another' solvent. In many cases, microwave digestion may be preferable because it destroys the sample matrix and allows a solution of sufficient sample concentration to be produced, enabling the laboratory to achieve the Method Reporting Limit and ensuring the maximum lifetimes from the plasma components that are exposed to the sample solution. In our view as a testing laboratory, it would be preferable to allow microwave digestion to be used as a first resort preparation technique, (where appropriate) rather than as a last resort when all other methods have proved ineffective.

The author says...

Conclusion

These critical observations of the stimuli article should only serve to illustrate the enormous complexity involved in creating pharmacopoeial methods that are required, in effect, to cover every current and foreseeable circumstance arising from pharmaceutical production. That it has taken some time for the revision of Chapter <231> to get to this stage should come as no surprise to anyone, and it is to be hoped that the final revision brings to an end what has been a lengthy and sometimes faltering process. An earlier attempt to revise the chapter had to be reversed and it would be a shame if the current process met with a similar end.

Reference

1. www.usp.org/pdf/EN/USPNF/2008-04-10InorganicImpuritiesStim.pdf