Header Artwork
Header Artwork


Chemical Characterization Has Come a Long Way

Prior to this past spring the latest version of the ISO 10993 Part 18 was more than 13 years old and was more of a descriptive standard highlighting what kind of techniques and approaches one could possibly take.  There was a lot of room for interpretation and misinterpretation on how to perform a Chemical Characterization study.

About 5 years ago, our sales person at our Leuven lab came to me and said, “I have this great opportunity with a medical device company to offer Extractables & Leachables testing.” “OK, great!” I said, and I thought that this was a very interesting development. We had been working on Pharma Extractables & Leachables (E&L) applications for more than a decade at that time, so I thought that this would give us new opportunities as a testing lab.

So, I started to put a testing program together, based upon the experiences and best practices that exist in Pharma E&L, but now translated them into testing a Medical Device. As the risk to the patient was more or less the same as for Pharma, I did not see the point in cutting back on any of the E&L scientific approaches.

A week later, the sales person came back to me and said, “We aren’t going to get the study; they are going for an approach that is only 10% of our price.”  I replied, “This simply cannot be possible, it is impossible to do all the things we are doing for 10% of the price.  We must be comparing apples to oranges here.”

I sent her back to consult again with the customer. A few moments later, she stepped into my office and said, “You were right, their approach is substantially different than ours. The methodology they will be using is to perform a Non-Volatile Residue (NVR) test and then perform a Fourier Transform Infrared (FTIR) analysis on the residue. Could we also offer this? That would bring our pricing more in line with the other lab. And this approach was accepted by the authorities!”

I did not know where to begin. We could not perform the NVR-FTIR any better or more accurately than anyone else; we could not do it more efficiently, nor could we do it cheaper. However, there was something more that was bothering me, “How in the world can you perform a sound and acceptable safety assessment using only on an NVR-FTIR evaluation?”

  • First, how relevant is an NVR? With a drying step at 105°C for 24 hours, you lose all of your volatile organic substances. These substances represent a class of compounds with the highest level of concern of all classes when it comes to mutagenicity and carcinogenicity. On top of that, you lose over 60% of your semi-volatiles as well. We have conducted the test, and observed that before and after the drying step, all of the semi-volatile compounds were gone from our chromatogram even with a retention time of less than 30 minutes (our total chromatogram is 45 minutes).
  • Next, how good is FTIR as a detection technique? While this test allows you to qualitatively indicate the rough composition of the residue (e.g. we see that amides and ketones are present) but it does not allow you to categorize the compounds to an identification and quantification level where their identity and dose could be linked to any relevant toxicological information.
  • Although both of these methods have purpose and scientific value; in this situation, the customer relying on NVR-FTIR to make a full E&L assessment of the device was not advisable. Too much information about the compounds for volatile and semi-volatiles found on the device is missing.

In one of my previous blog posts, I said we made all the mistakes one can make in Extractable & Leachable research. I need to rephrase this, all except one. We never performed an NVR-FTIR as a Chemical Characterization for Medical Devices. As a scientist, I would have lost my scientific integrity in doing so, and even if it would have contributed to lab revenue, I think in some cases it is more important to safeguard your reputation as a testing lab, than going for the bucks.

A few years after our first experience, we had another round of reality checks. The offering we were competing against had moved from NVR-FTIR towards GC/MS, which was a step in the right direction. However, to fully characterize an extract you need complementary and orthogonal chromatographic techniques—something we knew and had been practicing for over 15 years.

Can we blame these labs for their offering? I would definitely say no. When you are new in a certain field like E&L and Chemical Characterization, you try to find a way that makes sense from a scientific perspective and what does not. As I said many times before, we have been there too. However, it was clear that we had progressed further on the E&L learning curve than the other service providers at that time.  Because of the work we had been doing in Pharma E&L, our eyes were wide open, and we knew not to go in this direction.

Today I think that everyone knows what a true Medical Device E&L study needs – in terms of methodologies and instrumentation – to perform a sound Chemical Characterization assessment. You must assess the full spectrum of compounds on the device including volatiles (VOCs), semi-volatiles (SVOCs), non-volatiles (NVOCs), and metals.  You need at least Headspace GC/MS (VOC), GC/MS (SVOC), LC/MS (NVOC) and ICP (Metals), adding more specific methods when necessary. It is clear that the Medical Device Industry has gained a lot of maturity knowing how to perform a Chemical Characterization study on Materials and Devices properly. Common Pharma E&L practices have gradually migrated into Medical Device E&L testing—and have become accepted practices.

The optimization of the approaches now resides much more in the nitty-gritty details of how to optimize the methodologies, the proper preparation of samples, the execution of the analysis, the data interpretation, and the subsequent reporting. I think we are beyond the point where we need to raise fundamental issues regarding what type of information is needed to perform a safety assessment and what kind of methodologies and instruments could provide us with this information.