Can Cytotoxicity Findings Serve as Reliable Predictors of In Vivo Outcomes?

Biological risk assessment

Cytotoxicity testing is an integral part of the required biological evaluation for nearly all medical devices according to ISO 10993-1. Typically, medical device extracts are evaluated to estimate the biological impact of potential leachable substances that may be released from the device during clinical use. For extract-based cytotoxicity assays, extracts are prepared using cell culture media under exaggerated conditions (i.e., with agitation at elevated temperatures for extended periods) to enhance the identification of possible toxicological risks. Cytotoxicity describes the ability of a substance to inflict damage or result in the death of living cells. Cytotoxic agents can trigger various cellular responses including slowing cell growth, causing cell death, and interfering with the cell’s metabolism or structural integrity – but can the measurements of cell toxicity in a Petri dish (or multi-well plate) predict the overall biological safety of a medical device?

For decades, researchers have been working to determine if links between cytotoxicity and in vivo endpoints exist. The chemical industry was first to research the relationship between cytotoxicity and in vivo biological effects and began applying the apparent correlation to toxicity screening. In 2006, the National Toxicology Program (NTP) Interagency Center for the Evaluation of Alternative Toxicological Methods (NICEATM) and the European Centre for the Validation of Alternative Methods (ECVAM) conducted a validation study to characterize two in vitro cytotoxicity assays for determining starting doses for rodent acute oral toxicity studies. This work resulted in high quality in vivo lethality and in vitro cytotoxicity reference databases and refinement of LD50 values for many reference substances but ultimately demonstrated that cytotoxicity assays are not sufficiently accurate to correctly predict rodent acute oral toxicity. Cytotoxicity assays are prone to sometimes overpredict the toxicity of the least toxic substances and underpredict the toxicity of the most toxic substances. Why the apparent disconnect?

Classic cell assays capture straightforward cytotoxic effects but miss complex interactions. A leachable that might be diluted, metabolized, or excreted in vivo can damage a vulnerable cell monolayer in vitro. This sensitivity means false positives are common. Cytotoxicity assays often flag extracts that would never actually cause harm in a patient. At the same time, some biological reactions in vivo (such as an immune-mediated allergy) don’t manifest as cell death in vitro, leading to occasional false negatives – non-cytotoxic materials occasionally still trigger a sensitizing effect or cause systemic toxicity through other mechanisms which are not reflected in in vitro assays. Additional considerations include:

  • Minor test protocol variations (e.g., serum concentration, choice of medium, cell type, exposure method) influence cytotoxicity assay outcomes; in practice, ISO 10993-5 allows considerable leeway in its methods. Researchers have demonstrated that for the same material extract, cytotoxicity could range from negligible to severe depending on these conditions.
  • Interlaboratory variability is a well-known contributor. Since ISO 10993-5 was first published in 1992, it has already been revised twice to incorporate lessons learned from two separate interlaboratory studies conducted by the ISO committee. ISO’s Technical Report (ISO/TR 10993-55:2023) describes the results of the most recent interlaboratory study conducted with 12 participating laboratories. In this study, participants attempted to identify a non-cytotoxic material (a high-density polyethylene sheet), as well as two cytotoxic materials (segmented polyurethane films containing either: 0.1% zinc diethyldithiocarbamate “ZDEC” or 0.25% zinc dibutyldithiocarbamate “ZDBC”, both of which serve as positive controls in cytotoxicity assays). While all participants correctly categorized the materials, considerable variation of the results among the laboratories was detected. A more recent interlaboratory study across 52 international labs resulted in low reproducibility and determined that the specifications set by the current standard were insufficient to obtain comparable results for an identical medical device. Labs tested two samples: a non-cytotoxic polyethylene tubing and a polyvinyl chloride tubing, for which a cytotoxic potential was presumed due to the formulation and embedded plasticizer. Surprisingly, only 58% of the labs correctly classified the materials.
  • A living system can sometimes “detoxify” or buffer a material (via protein binding, absorption into tissue matrices, fluid dilution, etc.) in ways that a static lab extraction and simple assay cannot. A transient cytotoxic effect in vitro or in short-term implantation might not matter for long-term biocompatibility once the initial burst of leachates dissipates; for example, glutaraldehyde-fixed porcine tissues (grafts) may demonstrate a mild localized acute effect from the residual amounts of glutaraldehyde (a common crosslinking agent); however, over time it is neutralized by host proteins and the tissue grafts become relatively inert and provide a stable surface for host cell integration. Researchers have also demonstrated that polyurethane implants spiked with the known cytotoxin zincdiethyldithiocarbamate (ZDEC, frequently used as a positive control under in vitro conditions) cause a thicker fibrotic capsule and inflammation at one week, but the ZDEC is completely leached out and degraded by six weeks. When considering only an acute response, many clinically useful implant materials may be unnecessarily rejected due to the results of in vitro

In other words, a “cytotoxic” result in vitro does not mean a medical device is toxic in vivo, and conversely, a non-cytotoxic result doesn’t guarantee absence of irritation or other effects. Does this mean that cytotoxicity testing is useless? Not at all…it remains an indispensable screening tool that is necessary, but not sufficient, for evaluating biological safety. Current regulations and standards already reflect this understanding:

  • ISO 10993-5 advises that cytotoxicity testing results require expert interpretation. A “fail” does not automatically equate to an unsafe medical device. Factors such as the nature of the effect (slight vs severe reactivity), known device materials, and extraction conditions can help determine next steps.
  • ISO 10993-1 and FDA’s 2023 Guidance on Biocompatibility both endorse a weight-of-evidence approach. Rather than a checklist mentality, regulators expect a Biological Evaluation Plan/Report that integrates all data: cytotoxicity is just one piece of the puzzle.

In practice, when a cytotoxicity assay fails, next steps might include:

  • Investigate the Cause: As part of the risk assessment, identify what leachable chemical might be responsible for the cytotoxic response of the extracts. For instance, if an extract’s pH was very low/high or it contained an obvious cytotoxin, that provides a clue.
  • Assess Patient Exposure: Determine if the cytotoxic compound would reach patients at a harmful level. Many such failed cytotoxic extracts may contain only trace levels of compounds and do not present a risk to patients. Use tools like chemical characterization and literature with toxicology data to put the result in context.
  • Consider Re-Testing: The ISO 10993 series of standards allow using more “physiological” extraction conditions, if justified. For example, if a device is only in contact with blood for 1 hour, the standard 72-hour extraction in cell culture medium at 37°C might be overkill. A shorter or diluted extract would simulate actual exposure. Similarly, if a device is intended for brief skin contact, a mild extraction (or even a direct contact test on agar with limited exposure time) might be more relevant than an exaggerated extraction of the device.
  • Use Additional Tests Judiciously: If uncertainty remains, targeted testing can help. The field is evolving and new approach methodologies are emerging. Awareness of these can ensure better testing programs are designed and pitfalls are possibly avoided (or take advantage of new, more predictive assays as they become validated).

The key for regulators is documentation and expert analysis. In the Biological Evaluation Report, it must be explained how cytotoxicity failure was addressed: which chemicals were found, what their toxicological profiles are, and why, despite the in vitro result, the medical device is safe (or how the risk has been mitigated). If changes were made (e.g., added a rinse step to the cleaning process, changed a material, adjusted a sterilization method) to resolve cytotoxicity, those actions must be documented. Regulators do not automatically reject a device because of a cytotoxicity failure, but the scientific rationale provided for the result will be heavily scrutinized. The goal is to ensure patient safety with a minimum of unnecessary delays or tests.

MED Institute provides comprehensive biological risk assessment services to support the safety evaluation of medical devices in accordance with ISO 10993; our multidisciplinary team of professionals offers expert guidance on material characterization, biological equivalence assessment, toxicological evaluation of chemical constituents, and even support in responding to submission deficiencies. MED Institute technical experts have extensive experience in developing Biological Evaluation Plans/Reports in compliance with global regulatory standards and based on sound scientific principles. By engaging with our team of experts, several manufacturers have designed safer products and accelerated their regulatory submissions and successful market entry.

Contact us today to discuss your project needs 855.463.1633 | askmed@medinstitute.com | medinstitute.com.

OUR COMMITMENT

We are committed to consistently performing services with high quality, that deliver exceptional results, and add value to the client’s business.
For client surveys sent since 2025, we received ratings of 4.99/5 (16).