Process steps involving transition metal catalysts are now commonplace in API manufacture, presenting the real possibility for traces of these metals to remain in the API after purification. Common metals used in this way include chromium, copper, nickel, palladium, platinum, rhodium and platinum, although there are others. Historically, the most likely culprits for trace metals in drug products were arsenic, cadmium, lead and mercury, all of which are much more likely to enter the manufacturing chain from natural sources.

While the regulators are increasingly concerned about traces of heavy metals in pharma products, the testing protocol laid down in USP is extremely out of date, and the update that is being discussed and validated now is very much needed. The existing method, USP , is a purely qualitative test, involving sulfite precipitation of any metal that might be present before a visual comparison is made. This involves a color comparison, and not only is it difficult to spot very low levels of the metals, it gives no indication of the amount of metal that is present, or its identity.

Ten metals are specified in the test, but many of those all-important metal catalysts are omitted. The list in USP includes antimony, arsenic, bismuth, cadmium, copper, lead, mercury, molybdenum, silver and tin. Clearly this list is in serious need of an update, and a more up-to-date quantitative method would be a huge improvement.

New testing protocols are now available that involve inductively coupled plasma (ICP) techniques, notably ICP atomic emission spectroscopy, ICP optical emission spectroscopy and ICP mass spectrometry. ICP-MS is particularly useful for samples where the limits are lower as it has lower detection limits. Concentrations as low as 1ppb can be detected in this way, which is far beyond the capabilities of the old visual comparison technique. In addition, the fact that the electrons that are excited during the ICP process emit energy at specific wavelengths enables the exact nature of the metal contaminant to be pinpointed.

Two new protocols, USP and USP , are under discussion. These take advantage of the advances that have been made in detection technology, and expand the list of metals that must be tested for to reflect current concerns. Now, checks must be made for traces of arsenic, cadmium, lead and mercury in all samples, plus chromium, copper, iridium, molybdenum, nickel, osmium, palladium, platinum, rhodium, ruthenium and vanadium where appropriate. Although they have yet to be agreed, approved, adopted and implemented, the industry is already working on how they will be implemented at a practical level.

One of the aspects that must be determined ahead of implementation is the development and validation of methods for sample preparation. There are four ways in which samples can be prepared, depending on the nature of the drug product. If the substance is a liquid, then it can be tested neat and, similarly, if the drug product is formulated as a solution for injection or infusion, this can be used directly. Otherwise, a solution will need to be made. If the solid is water-soluble, a direct aqueous solution would be appropriate; similarly, if it will dissolve in a suitable organic solvent, a direct organic solution can be used. However, an aqueous solution is preferable, as the organic solvent can be harsh on the instrument, and increase the difficulties involved in identifying the optimal analytical parameters for the test.

The worst-case scenario is that an indirect sample must be made. This is typically done via microwave digestion within a closed vessel, a process that breaks the insoluble components into smaller fragments that are soluble. This process takes time; it’s not as simple as simply adding sample and solvent to the closed vessel and putting it in the microwave. Rather, the best acid and heat profile for the digestion has to be determined, and the lowest temperature and shortest time in which the digestion occurs. Each drug sample will be different in terms of its digestion behavior, and developing the protocol for each is somewhat laborious. However, in the absence of a direct sample preparation method, it must be done.

Direct aqueous solutions: not always straightforward
One might imagine that if the product will dissolve in water, there will be no problems in running the test. In practice, this is not necessarily the case. The spiking studies for validation on one such example, a sample that was provided as a ready-formulated solution for infusion, provides an object lesson in why one should never make this assumption.

The validation studies required three samples to be prepared for the different metals. Sample one included 2% nitric acid in order to make it suitable for spiking with cadmium, chromium, lead, mercury, molybdenum, palladium, ruthenium, silver and vanadium. Sample two needed a stronger acid solution, with 12% nitric and 8% hydrochloric, for spiking with copper, iridium, nickel, platinum and rhodium. Sample three included 2% hydrochloric acid to make it suitable for spiking with osmium.

All three samples were then spiked with the appropriate metals, each at three different levels of spiking, before being tested. USP accuracy requirements were met for the majority of these elements, but several fell short, with copper, nickel, osmium, platinum and rhodium failing to reach the required 70% average recovery.

It was clear that something untoward was going on in the sample with these five metals. Therefore, the sample was subjected to microwave digestion in the hope it would break the drug product down into smaller fragments that did not interfere with the detection of the metals. This was largely successful, with the exception of osmium, a metal we had previously had difficulties with.
So, rather than microwave digestion, the infusion solution itself was used, with hydrochloric acid as the diluent rather than nitric acid. This time, acceptable recovery of osmium achieved. The problems experienced here exemplify the importance of running careful validation studies for each new drug product, however straightforward it might appear to be at first sight.

Purification matters
It is vital that validation studies are carried out on the exact same type of material that was used to develop the sample preparation method, otherwise unforeseen problems can occur. For example, a customer supplied a purified form of the material for the validation studies, whereas to save time and money the material that had been sent for sample preparation development had been unpurified. This in itself is not necessarily a problem, but here the samples suddenly changed from passing on all criteria in the development work to running low for palladium.

The purification had involved the use of acid, which left the material more acidic than the unpurified version had been. For some unknown reason, this had an effect on palladium alone, with all the other metals behaving in the same way as they had in the unpurified material. Palladium is one of the most important metals to get right, as the number of palladium-based catalysts that are in use in the synthesis of active pharmaceutical ingredients is significant, and palladium cannot be identified via the old color comparison tests.

The solution to this problem lay in changing the dilution factor. In the original sample preparation, the concentration had been 1.15µg/g. By reducing the dilution slightly and using a sample solution of 1.25 µg/g, with a sample volume of 25ml instead of 10ml, an acceptable palladium recovery was achieved from the spiked samples. Clearly, the sample amount and dilution factor were both critical in being able to recover palladium at low levels. Curiously, the recoveries before purification did not seem to be as affected by changing the dilution factor.

Choosing the correct digestion method
When digestion has to be used because none of the weigh and dilute techniques are appropriate, it is important to develop a digestion method that gives reliable and reproducible results. In this example, a digestion protocol to prepare the sample had already been developed by the customer. It involved the addition of nitric acid and hydrogen peroxide, before heating to 105°C in a regular oven. This is faster and simpler than using microwave digestion, and had proved successful on all samples throughout the product development process. But, as in the previous example, palladium recovery came out low when the purified product was used in the validation work.

This time, there was no simple fix, and the only solution was to start again, with the aim of creating a whole new sample preparation method that would work for the entire panel of metals, rather than coming up with a new one just to fix the palladium problem. While this is not always possible, avoiding the need to prepare different samples for different metals will clearly save time and money.

In order to minimize the possibility of negatively impacting the recovery of all the other metals, the starting point was the same combination of nitric acid and hydrogen peroxide that had been used in the initial method. Simply switching from normal oven heating to microwave heating was sufficient to create samples that gave acceptable recovery across the board, including palladium. The validation process was successful, and this method is now routinely used for preparing test samples of this product.

When similar methods do work
If the same method can be applied to multiple samples, it can greatly speed up development times. Of course, this does not always work, and even when it does, the same careful validation procedures must be applied to ensure that all is well and the sample preparation method is indeed appropriate in all cases.

Here, the development was done on a free base sample for arsenic, cadmium, lead, mercury and nickel, and all the criteria were met. The tests were run on 0.1g of sample, with 2.5ml of nitric acid added before being microwaved, and then diluted to 12.5ml with water. Exactly the same method was applied to the hydrochloride form of the sample for those that required hydrochloric acid. Again, all the recovery requirements were met for these samples. While validation is still necessary in all cases, this does illustrate that it may be possible to apply the same sample preparation methodology to different product samples without having a major effect on recoveries. It is always sensible to try an already validated method on a similar product to see if it would work before spending the time developing something completely new.

Looking to the future
USP and USP have still not been adopted and are not yet widely used, although the two general chapters have been official since February 2013. A few monographs now reference them, so they are official, but USP has not been removed, and is still, for the most part, being used.

The full rollout of the new chapters has been further delayed to enable harmonization work to be carried out in conjunction with ICH and EP, with a projected date of December 2015. This is a sensible move, but it is making for an even longer and more drawn out process. However, it does give plenty of time for reliable and reproducible sample preparation methods to be developed that will make for much more rapid compliance with the new rules once they are finally implemented.

Although the documentation has yet to be finalized, it will not include any sample preparation methods. This is a good thing, bearing in mind the fact that sample preparation is rarely a one-size-fits-all procedure, with careful development required for each different drug product to ensure appropriate and adequate recoveries are achieved in the spiked samples, giving confidence that traces of heavy metals will not be missed in real-world samples. CP

---

Nikki Schopp received her BS in Biology with a minor in Chemistry, Marine Science and Psychology from The University of Miami in Florida in 2005. Ms. Schopp joined SGS in 2005 and has been concentrating on ICP since 2007. In her current role, Ms. Schopp is the Team Leader in charge of ICP-MS, ICP-OES and AA testing. Recently her focus has been on elemental impurities method development and validation.