The Hippocratic Oath, attributed to the Greek physician Hippocrates, required doctors of the time to “abstain from whatever is deleterious and mischievous,” to “give no deadly medicine,” and otherwise avoid any voluntary act that may cause harm. The maxim “first, do no harm” is part of a broader set of commitments that modern doctors promise, and all who contribute to the art of healing should abide.

Indeed, within the medical device industry, we seek to enhance the lives of patients through technology and design and must never bring a device to market that may do more harm than good. There are many ways in which a medical device could inadvertently cause harm, confirming that a device doesn’t cause an unacceptable risk of harm amounts to a demonstration of “biocompatibility.”

Verifying that a device is biocompatible is a necessary burden. Of course, devices are designed to be biocompatible, but history has proven that unforeseen complications or contaminants can cause devastating harm.

As case studies rise to the surface where people have been harmed, regulatory scrutiny is renewed and the strategy to prove biocompatibility changes. Prior to the turn of the millennium, evaluation was very much a check-box approach. The G95 included a table of “Initial Evaluation Tests for Consideration,” that would inform the entire biological evaluation plan. Knowing the contact type and duration of contact, a manufacturer simply ordered the required tests, and if they all passed, the device was biocompatible. The problem with check boxing is that it required no knowledge of device materials and any special risks those materials would present. The outcome was that very often burdensome tests were executed and carried out unnecessarily on materials certifiably biocompatible, and other times materials were used, like fishing line from the sporting goods store, which could technically pass the test but do not belong permanently in the body.

Therefore, ISO 10993-1 was drafted and revised to insist on use of an intelligent and risk-based approach. Knowledge of the device materials and how those materials may interact with the body was made a prerequisite to conducting animal tests. The FDA issued guidance on application of a risk-based and least-burdensome approach. Abandonment of check-boxing necessitated the formulation of strategies to effectively evaluate biocompatibility: written documents that lay out the materials, associated risks, and how those risks would be tested. The Biological Evaluation Plan (BEP) has become the frontline vehicle to summarize this information (see Figure 1). Following formulation of the BEP, tests and other scientific justifications are made, and the results consolidated and summarized in a capstone story called the Biological Evaluation Report (BER).

Regulatory expectations for the contents and level of rigor in a BEP and BER have rapidly been changing since the latest update to ISO 10993-1:2018, ISO 10993-18:2020, and the European Regulation 2017/745 (EU MDR). One key requirement of 10993-1:2018 is that all devices, regardless of patient contact or perceived risk, should have “physical and/or chemical information.” While it is explicitly clear that “physical and/or chemical information” does not mean chemistry testing, many device manufacturers have pursued extractables chemistry testing for the first time as a result of this requirement.

Caution must be exercised in deciding to do extractables testing, as not every device type is amenable to such an analysis, and there are examples where generating extractables data can cause more harm than good in a submission. The first question should be whether chemistry testing is truly necessary; if the device is limited contact or resorbable, other options should be explored first. When chemistry testing is necessary, the parameters must be perfectly clear and checked against the most recent regulatory expectation.

The requirements have shifted dramatically over just a few short quarters. Remember, the only reason to conduct chemistry testing is to mitigate biological risk that would otherwise require an animal test. Therefore, the chemistry results need to be reviewed by a toxicologist who will make a claim regarding safety. The parameters of testing must also be reviewed by the toxicologist who will make the claim (toxicologists have very specific requirements on data quality).

Shifting Requirements for Chemical Characterization

The most significant shifts in strategy with regard to chemical characterization pertain to two key points: the sensitivity of analysis and required system suitability information. The required sensitivity is the so-called analytical evaluation threshold (AET), a term that became widely accepted after the 2020 published update to 10993-18. What is critical today is the manner in which the AET is established.

The FDA expects a dose-based threshold to be the foundation of the AET, divided by an uncertainty factor that ensures that chemicals observed but underreported are captured for assessment. Importantly, the FDA accepts only one formula for AET while there are a couple of options published in 10993-18, and the uncertainty factors have certain requirements (such as being supported by labspecific data). The newly required system suitability data is more variable, but in general includes a minimum number of fully validated surrogates for each method, across a spectrum of chemical properties, where it is proven that the sensitivity of analysis for these surrogates is at least as good as the proposed AET.

Another recent system suitability requirement is the need for spike and recovery data across sample preparation steps that often include liquid-to-liquid solvent transfers and concentration steps. All of the following main parameters should be specified as part of the BEP: extraction time and temperature, solvents to be used, and AET.

Shifting Requirements Regarding Shelf-Life

During the height of the COVID-19 pandemic, there was an acute shortage of N95 masks for use in healthcare settings. Work was happening at a breakneck pace to attempt validation of processes to clean masks for reuse while manufacturers ramped up production. Nurses were wearing old and soiled masks, wearing unvalidated masks, or making their own to give themselves some level of protection.

The Hippocratic Oath, attributed to the Greek physician Hippocrates, required doctors of the time to “abstain from whatever is deleterious and mischievous,” to “give no deadly medicine,” and otherwise avoid any voluntary act that may cause harm.

Against this backdrop, a large manufacturer of N95 masks found several hundred thousand masks in a warehouse. These particular masks had a shelf life of four years, but had sat on the shelf eight years. Due to the gravity of the need, the masks were sent to hospitals for use. Many users of the masks made complaints of respiratory irritation. Of course, use was paused, and the manufacturer reviewed the historical biocompatibility of the devices. It was clear that they had passed all required biocompatibility.

Upon review by the Expert Advisory Services team at Nelson Labs, it was found that these particular masks had a polyester foam strip along the nose piece to help prevent breath from fogging the user’s glasses. Polyester foam can slowly degrade and particulate, producing respirable particles that in the environment of the lungs further hydrolyses into mildly acidic molecules known to cause temporary respiratory irritation. The remaining masks were destroyed.

The mask case study described is an excellent example of a situation weighing on the mind of FDA: how do we know whether the biocompatibility of a device slowly changes for the worse on the shelf? End-of-shelf-life testing hasn’t been a requirement historically, but there are examples of patient harm resulting in recalls due to this. Therefore, end of shelf-life biocompatibility should, at a minimum, be directly discussed in a BEP.

Most often, this discussion is short and doesn’t require any additional testing. “We know that Ti6Al4V is a material extraordinarily resistant to any degradation or corrosion under conditions extremely exaggerated compared to those used in storage; therefore, the biocompatibility at end of shelf life is expected to be unchanged.” However, sometimes, upon review it is clear that a limited amount of testing such as cytotoxicity and comparative FTIR is warranted to establish material stability.


Modern evaluation of biocompatibility uses a risk-based approach that requires finesse. The testing burden is generally less than a decade ago, but much more attention and effort is needed in the formulation of an appropriate strategy. For most devices, expert knowledge of materials is needed, along with feedback from a toxicologist, and an up-to date pulse on expectations of regulators. BEPs have expanded from 1–2-page documents to 20 pages long, a reflection of the thoughtfulness required as doctors, device manufacturers, expert advisors, and labs partner to ensure that what makes it to market “first, does no harm.”

This article was written by Matthew R. Jorgensen, PhD, DABT, Chemist, Materials Scientist, and Toxicologist for Nelson Labs, Salt Lake City, UT. For more information, visit here .