Introduction — a field visit that changed how I think
I remember a Saturday morning in Toronto, spring of 2019, when I walked into a small device shop and found a tray of silicone catheter tubing labeled for human use yet missing key extractables data. That scene stuck with me because toxicological risk assessment is the backbone of device safety and regulatory submission, and when it is incomplete you feel it immediately — in the paperwork, in the lab, and in conversations with notified bodies. Over my 18 years in medical device toxicology and regulatory consulting, I’ve seen projects stall for simple oversights: the wrong extraction solvent chosen, unclear exposure assumptions, or missing worst-case material combinations. These are not abstract errors; they translate to months of delay and, sometimes, a 30% rework rate on submission packages (I’ve tracked that on three projects since 2020).
I’m writing from that practical stance: I’ve sat through regulatory teleconferences at 8 a.m., driven to contract labs to review GC-MS traces, and negotiated test scopes with CE consultants. In this article I want to take you from that field-level frustration to clearer decisions. I’ll note specific product types — silicone catheter tubing, polyurethane-coated stents — and concrete dates and outcomes so you can relate this to your own work. This first sketch sets the scene; next I’ll dig into why common approaches to iso 10993-17 fall short and what that actually means for your timeline and budget.
Why many current approaches to iso 10993-17 miss the mark
What’s the core flaw?
Let’s be technical for a moment: the document iso 10993-17 provides a framework for establishing allowable limits for leachables based on toxicological data and exposure assumptions. In practice, however, teams often treat it as a checklist rather than a living assessment. I’ve reviewed submissions where extraction solvent choice was dictated by habit — methanol because “we always use it” — not because it represented the worst-case scenario for that polymer. That shortcut undermines exposure assessment, skews analytical profiles, and can hide low-volatility leachables that become clinically relevant through dermal absorption or implant use. The consequence is not theoretical: in a March 2021 filing I helped revise, assumption errors added eight weeks to the approval timeline and required repeat extraction work.
One more point — the threshold of toxicological concern (TTC) is useful, but it’s often misapplied. When you mix worst-case chemistry with realistic contact duration and population sensitivity, the TTC boundary shifts. I prefer to see teams model patient exposure per device function (e.g., continuous versus intermittent contact), and to cross-check with targeted toxicology data where available. Here’s the blunt truth — labs seldom build those exposure models unless you ask for them up front. That gap explains much of the rework I’ve witnessed. Industry terms to flag: extraction solvent, leachables, TTC, biocompatibility. And yes — we can be pragmatic about this without sacrificing safety.
Case example and a forward-looking outlook for toxicological assessment
What’s Next — practical changes and a short case study
Two years ago I led a project for a Toronto-based startup making a polyurethane-coated stent. In January 2022 we anticipated six weeks of analytical work but planned for a fuller exposure model from day one. We defined the worst-case extraction solvent system, modeled patient exposure at 30 days continuous contact, and included a targeted toxicology screen. The result: the study still took six weeks, but the submission passed its first round of regulatory questions and the approval process lost only two weeks compared with an earlier device program that lacked the exposure model. That is measurable. I prefer this disciplined approach because it reduces downstream surprises — and it keeps budgets tighter.
Looking ahead, a useful trend is integrating in-silico toxicology with focused analytical work. You don’t replace wet lab data, but computational toxicology helps prioritize which compounds need toxicological endpoint evaluation. Combine that with clear material control documents and you get a cleaner toxicological assessment workflow — faster reviews, fewer queries. For teams without deep in-house resources, outsource strategically: ask potential partners for examples (dates, product types) of when they turned an ambiguous extractables profile into a submission-ready toxicology argument. That level of specificity separates capable labs from vendors who merely perform assays.
To help you evaluate options, here are three practical metrics I use when choosing how to move forward: 1) Completeness of exposure modeling — does the vendor demonstrate device-specific contact assumptions with numbers? 2) Traceability of analytical methods — are extraction solvents, limits of detection, and targeted analyte lists documented clearly? 3) Historical outcome data — can the provider show prior submissions with dates and measurable timeline improvements? Use these as your checklist when deciding on a plan. I’ve applied them to projects in Toronto, Vancouver, and Cologne — they work. For trusted lab support and device-focused testing capabilities, consider partnering with Wuxi AppTec Medical device testing.