A Practical Guide to Toxicological Risk Assessment for Medical Devices

Medical device manufacturers operate in a different regulatory environment than even a few years ago. Expectations have shifted from demonstrating a device passing a set of biological tests to showing a clear, science-based understanding of patient risk.

At that shift’s center is toxicological risk assessment (TRA). Once viewed as a supporting activity, TRA now drives how manufacturers interpret chemical data, justify safety, and respond to regulatory scrutiny. It connects extractables and leachables (E&L) data to real-world patient exposure and translates that information into a defensible risk assessment.

The challenge is understanding how the core components of a TRA fit together. We still see companies approaching hazard identification, exposure estimation, and risk assessment as separate tasks. In reality, they work together to tell one cohesive story about patient safety.

Historically, biocompatibility followed a predictable path. Manufacturers categorized a device, selected tests from established matrices, and generated data for submission. Passing those tests often signaled the device was safe.

That approach no longer holds up on its own. Recent updates to ISO 10993 reinforce a risk-based framework that emphasizes understanding over checklist execution. Regulators expect manufacturers to explain what patients may encounter, quantify the level of exposure, and determine whether that exposure presents a meaningful risk.

TRA enables that explanation. TRA evaluates chemical characterization data, identifies potential hazards, estimates patient exposure, and determines whether those exposures fall within acceptable limits. When teams do TRA well, the process grounds decisions in evidence and closes gaps that testing alone can’t cover.

This shift also changes how manufacturers respond to unexpected results. A failed biological test prompts more analysis, often including chemical characterization and a structured risk assessment to identify the cause and determine next steps.

Every TRA starts with hazard identification and asks: what could go wrong? For medical devices, that means examining base raw materials, manufacturing residues, processing aids, degradation byproducts, and compounds identified through E&L studies.

Device categorization plays a central role. An implantable device presents a different risk profile than one with limited skin contact. The nature and duration of contact determine which biological effects must be evaluated, from irritation and sensitization to systemic toxicity, genotoxicity, carcinogenicity, reproductive effects, and developmental toxicity. Manufacturers typically leverage safety data sheets for materials and processing chemicals, published literature and toxicological databases, prior knowledge of similar materials or devices, and E&L data generated during testing to identify hazards.

A common mistake at this stage is equating detection with danger. Analytical methods continue to improve, enabling detection of trace levels of compounds that would’ve gone unnoticed in the past. Detection alone doesn’t indicate risk.

Manufacturers must distinguish between what’s measurable and what’s meaningful, which requires setting appropriate analytical thresholds, and understanding how those thresholds relate to toxicological relevance.

Another frequent challenge is misinterpreting worst-case scenario data. Aggressive extraction conditions can reveal compounds that would never appear under normal use. Without careful interpretation, these results can lead to overestimating hazards. At the same time, incomplete data or a cursory toxicological review can lead to underestimation. Both situations create issues due to unnecessary testing or missed safety concerns.

Hazard identification focuses on what could cause harm. Exposure estimation looks at whether patients will encounter it and how much they’re exposed to. This step relies heavily on E&L studies, which simulate what may migrate from a device under controlled conditions. These studies often use elevated temperatures, extended durations, and multiple solvents to create a worst-case scenario.

Worst-case exposure is a conservative baseline to help keep potential hazards from slipping through the cracks in a risk assessment. However, real-world exposure rarely matches testing conditions. Medical devices are used at body temperature and follow defined exposure paths. Duration of contact with the device (the biggest factor influencing exposure), frequency of use, surface area and design characteristics, stability of materials under use conditions, and extraction parameters used during testing shape exposure estimates.

Be cautious with assumptions; they often lead to errors. Assuming aggressive extraction conditions are needed can elevate exposure estimates over what’s realistic for patient experiences. We may also underestimate a device’s impact on pediatric patients, where lower body weight can change calculations. Manufacturers often rely on conservative assumptions when real-world data is limited. That approach can support the assessment but it’s important to explain those assumptions clearly.

One of the biggest biocompatibility misconceptions is treating hazard and risk as the same thing. Hazard is about a substance’s potential to cause harm; risk depends on both the hazard and the amount of exposure. A substance can be highly hazardous but pose little real risk if exposure is minimal. Conversely, even a lower-hazard substance can become a concern when exposure levels are high.

Consider stainless steel devices: Cobalt is a residue in stainless steel and is listed as a carcinogen. However, if the E&L study demonstrates cobalt doesn’t leach out from the material under use conditions, the presence of cobalt could be considered a hazard without exposure and patient risk.

Make individual assessments for each substance found to help eliminate zero-tolerance thinking. Not every detected compound will be a problem. TRAs help identify which compounds require more evaluation.

The margin of safety is the ratio between tolerable intake and worst-case estimated exposure. It provides a quantitative measure of how close a given exposure is to a harmful level. A margin of safety above one indicates that exposure remains below the level of concern. Values below one demonstrate the need for closer evaluation, potential risk mitigation, or design changes.

Uncertainty factors account for differences between study conditions and real-world use. They address variability between species, differences in exposure routes, and gaps in available data. These factors can significantly influence the final margin of safety. For example, extrapolating data from animal studies to humans introduces additional uncertainty, which must be reflected in the calculation. 

Certain populations may require more conservative assessments. Neonates and infants, with lower body weights and developing systems, often result in smaller margins of safety. In these cases, manufacturers may need to conduct a risk-benefit analysis to justify device use.

Margins of safety also help reduce reliance on animal testing. When chemical characterization and exposure assessments demonstrate acceptable margins, manufacturers may address certain biological effects without additional in vivo studies.

Risk characterization integrates hazard identification, exposure estimation, and margin of safety into a cohesive conclusion.

The process follows a structured path:

1. Identify potential hazards based on materials, processes, and analytical data
2. Estimate patient exposure using E&L data
3. Calculate margins of safety for each relevant substance
4. Evaluate whether exposures fall within acceptable limits

A strong risk characterization incorporates a critical review of the underlying data, including the quality of toxicological studies, relevance of exposure scenarios, and appropriateness of assumptions. It also considers patient populations, intended use, and foreseeable misuse. Assessments must address scenarios where clinicians use devices beyond their primary indication.

Manufacturers often underestimate the importance of considering all population groups. Children are often under-represented. Pregnant or breastfeeding women can also be under-represented. Rushed assessments or incomplete consideration of populations can lead to gaps that regulators quickly identify.

At the same time, overly conservative approaches add unnecessary complexity. For example, using the wrong solvents during extraction can produce results that look concerning but don’t reflect the device’s use in the real world. Data must be reviewed with the right level of balance.

TRA is a technical exercise and a communication tool. Regulators expect manufacturers to present a clear narrative explaining what they found, what it means, and why the device is safe for its intended use. This approach requires teams to translate complex data into a structured, evidence-supported argument.

Manufacturers must demonstrate how they will address identified risks. Options may include modifying materials or design, adjusting manufacturing processes, implementing additional cleaning or flushing steps, and conducting targeted testing to resolve uncertainties.

Risk-benefit analysis may also play a role, particularly for devices used in vulnerable populations or critical medical situations. The goal isn’t to eliminate all uncertainty but demonstrate risks are understood, controlled, and justified in the context of patient benefit.

Weak or incomplete TRAs carry real consequences. Repeat testing, delayed submissions, and extended regulatory reviews can add months to timelines. Sometimes manufacturers must redo studies, resulting in additional costs.

Regulators may raise questions requiring additional documents or analysis. Every round of questions extends the timeline and increases costs. Poor risk assessments can also undermine confidence in a submission. A fragmented or unclear narrative signals the manufacturer doesn’t fully understand the device’s risk profile.

A TRA should be performed as early as possible in the assessment to account for all chemical hazards. If started too late, there may be gaps in assessment or failed tests that cannot be explained. The growing importance of TRA reflects a broader shift in the industry’s evaluation of medical device safety.

Here are three key takeaways:

1. A robust TRA strengthens submissions by providing clear, science-based evidence of patient safety.
2. Testing strategies must prioritize quality over quantity. Evidence is essential. More data only matters if it’s meaningful.
3. TRA is the foundation of biological evaluation, with in vivo testing serving a complementary, not primary, role.

Manufacturers who embrace TRA gain insight into their devices, confidence in their decisions, and a stronger foundation for bringing safe products to market.

MORE FROM THIS AUTHOR: Why Toxicology Is Becoming the Backbone of Biocompatibility

Dr. Isabel Groh, ERT, DABT, is a senior toxicologist and biocompatibility expert at Hohenstein Medical, where she leads toxicological and biocompatibility evaluations and supports strategic development of Hohenstein’s medical device safety services. With dual board certifications as a European Registered Toxicologist (ERT) and Diplomate of the American Board of Toxicology (DABT), Dr. Groh brings extensive experience in toxicology, biocompatibility, and pharmacology.