USP's DeStefano on Defining, and Detecting, Pharmaceutical Impurities

Pharmaceutical impurities aren’t as simple as they used to be. 

Today, instead of having to worry about traces of metals from catalysts, ingredients or equipment, manufacturers must also consider the potential of economically motivated adulteration, and grapple with the complexity of different forms of complex biopharmaceutical molecules. 

Defining safe exposure limits, improving testing methods, and even deciding what an impurity is, are more challenging than ever, especially as scientifically-astute criminals engage in tampering for profit.

At USP, Patrick Lukulay [director of the USP-US AID program, Promoting the Quality of Medicines] and his team have been working on a database that could eventually serve as a resource for regulatory bodies, not only in the developing nations but globally.  These efforts tie in with FDA’s recent announcement on its collaboration on drug anticounterfeiting with the World Health Organization (WHO).  USP expects to be able to make an announcement on this early next year.

In the meantime, USP is working to modernize monographs and testing methods. For elemental impurities, the Pharmacopeia is in the process of updating its 105-year-old elementals impurities test.  In concert with ICH, the U.S. Pharmacopeia is also working to set acceptable metal exposure limits.

On his way to an ICH Q3D working group meeting in Japan this week, USP’s Vice President for General Chapters, Anthony DeStefano, described the Pharmacopeia's plans and some of the issues involved. 

He will return the following week for a special workshop on November 13-14 in New Orleans, “Impurities, Adulteration and the Changing Role of the USP in Global Drug Quality,” which will precede the annual AAPS conference.

PhM: When did USP first realize that the entire area of impurities was becoming such a problematic one? 

ADS: Impurities are a major problem on many levels. There are the elemental impurities, then the contaminants and adulterants, such as diethylene glycol (DEG), not only in drugs and APIs, but in dietary supplements, traditional medicines and biologicals. 

USP’s test for elemental impurities, for metals, first appeared in the USP in 1905.  We knew that that test needed to be upgraded.  Going back to 1995, we’ve had articles in our comment journal, the Pharmacopeial Forum, suggesting that they be revised, so this initiative has really picked up over the past few years. 

We’ve proposed toxicologically-based limits for metals that are known to be toxic, as either catalysts or contaminants. 

The ICH Q3D working group has picked up on this issue as well. 

USP has proposed methodologies for how to get to suggested levels because the current methodology is not viable.  But this methodology is focusing on addressing public safety, rather than exploring detection limits and “how little can we see?”

PhM: Are there specific detection technologies that have been proposed?

ADS: For metals, we are proposing primarily inductively-coupled plasma (ICP) with optical emission spectroscopy and ICP/MS as the primary detection tools, but other tools including atomic absorption spectroscopy can also be used. 

We’re going to propose some acceptance criteria for precision, accuracy and selectivity at the proposed limits of permissible daily exposure (PDE). 

As long as you meet these precision, accuracy and specificity criteria, you can choose a method that works.  For instance, if you don’t like ICP and you want to use atomic absorption with, say, a platinum lamp because you only need to look for platinum, you’ll be able to do that.

In general, we’re moving away from visual tests to instrumental tests.  The current USP elemental impurities test is precipitation followed by visual observation, which is very difficult to do with any level of accuracy or precision.

We’re working with ICH to see whether we can agree on which metals should be limited, and what the proper PDE levels should be.

PhM: Is there a timeline for setting PDE’s?

ADS: We hope to come out after this meeting with a Stage 2 draft, so that we’d have settled on which metals and which limits to focus on.  We’ll then put them out for initial public comment over the next few months. 

USP has also drafted some relevant chapters which we hope to put out for comment in the spring, and we expect these efforts to dovetail with those of ICH.  We’re assuming that all will go well and we’ll have a good idea of metals and limits over the next three months.

PhM: What are USP’s plans for biopharma impurity testing?

ADS: The biggest problem on the large molecule front has been heparin contaminated with  oversulfated chondroiten sulfate (OCS). The Pharmacopeia is really designed to characterize what is there, rather than find what is not there. 

So, for example, with DEG, as was found in the glycerin case, we have to test the identity of something that isn’t supposed to be there. So the identity is in part defined by something’s absence. 

Similarly, for heparin, there’s a test for level of OSC, and part of the heparin identity test is absence of OSC.  That’s done deliberately because of the way that GMPs are set up.  When materials come into a plant, GMPs require material identity tests.  So, we can only discover that something is missing during identity testing.  The other tests can be done any time.  When the truck is out there, at the door, you have to be able to answer the question, “Do I want this shipment or not?”  That’s done through identity testing.

But heparin is just the tip of the iceberg.  There are all these other biologics, which beg the questions: How does one characterize them, and what do you even call impurities in a biologic?  Let’s say you are dealing with a drug that is heavily glycosylated….are all those levels of glycosylation different forms of the drug or aren’t they? 

That question then spills over into the issue of biosimilars:  What is “close enough”?

A less glycosylated species that is active is still active, yet, technically, you can’t call that an impurity because it’s just another form of the drug. 

All these issues wrap around each other as you look at a complex molecule that is heavily glycosylated or that could have different conjugations. How do you define impurities with these types of compounds? 
That remains to be defined.

PhM: Does USP have feelers out in the field, which would allow industry to better anticipate what tomorrow’s adulterants might be, or is that just too difficult to accomplish today?

ADS: There’s a philosophical debate about how to approach this problem.  FDA has given USP  a list of other sugars like sorbitol, for which they want equivalent tests for contaminants such as  DEG.  The European and Japanese Pharmacopeia’s have, instead, taken the position that, so far, nobody has tried to contaminate these things, so equivalent tests aren’t needed.

 FDA’s approach is more proactive, as in “Let’s  do something before someone does,”  where EP and JP don’t want to give the impression that they’ve protected these molecules when all they would have done is look for one or two things. 

However, USP is continuing to approach this problem by developing more specific tests.  Consider melamine.  Instead of the traditional “total nitrogen” test, which was unspecific, we are moving to highly specific quantitative test.  That way, if I look for an active ingredient and find it’s 80%, I can reject a lot of product.  At this point, I don’t need to know what the other 20% is….I now know enough to say, when the truck is at the factory door, “I don’t want to use this shipment.”

The idea is not to try to track down every possible impurity, but to have a sufficiently specific and quantitative that can identify a problem with the active ingredient.  That’s the long-term strategy that has the best chance of success.

PhM: Is spectroscopy the preferred method for doing this?

ADS: Raman and IR Spectroscopy works well, but you can still use HPLC. 

If you use better and better HPLC, you can assure yourself that there’s nothing else under the main component. 

You may have to combine a couple of technologies.  For example, you could combine HPLC with Raman for the identity test. In this case, HPLC would tell you, say, that the active was greater than 95%.

Some things will depend on how we use spectral libraries.  Maybe, one can specify a certain hit ratio with a spectral library. 

What is clear is that specifying any one technology is not the way to go. Once today’s criminals figure out how you’re trying to deal with them, they turn around and figure out how to deal with you.  If you think about OSC, it was terrifically clever to be able to figure out a way to fool the biological assay test….and with something that is also very cheap. 

We’re dealing with a much more sophisticated bad person than we’ve had to deal with in the past, when guys were simply dumping impurities.  When heparin goes for $1200 and they’re making it for $20, there’s motivation.

In the end, as a colleague recently said, the person that will first detect adulterants is not your scientist, but your purchasing guy.  When you see the price of heparin skyrocketing and you get a really good deal on it, that may a signal that it’s time to start looking for a supply chain problem.

PhM: Are there any efforts to collaborate on developing spectral libraries?  I’ve heard of some efforts going on with Raman.

ADS: These efforts are at a very early stage. We do have a melamine working group with FDA.  Melamine is really the short-term case study, but we plan to broaden that work to develop screening methods for adulterants. Bill Cook [USP’s chief metrology officer] has a CRADA with the Agency, and one focus is on “dividing and conquering,” so that we don’t duplicate efforts.

We can’t do everything, so the critical questions were: Where do we start, and what do we protect first?  That’s why we started with melamine, but eventually we plan to expand efforts to other materials.