BACKGROUND: It has been suggested that carbon nanotubes might conform to the fibre pathogenicity paradigm that explains the toxicities of asbestos and other fibres on a continuum based on length, aspect ratio and biopersistence. Some types of carbon nanotubes satisfy the first two aspects of the fibre paradigm but only recently has their biopersistence begun to be investigated. Biopersistence is complex and requires in vivo testing and analysis. However durability, the chemical mimicking of the process of fibre dissolution using in vitro treatment, is closely related to biopersistence and more readily determined. Here, we describe an experimental process to determine the durability of four types of carbon nanotubes in simulated biological fluid (Gambles solution), and their subsequent pathogenicity in vivo using a mouse model sensitive to inflammogenic effects of fibres. The in vitro and in vivo results were compared with well-characterised glass wool and asbestos fibre controls. RESULTS: After incubation for up to 24 weeks in Gambles solution, our control fibres were recovered at percentages consistent with their known in vitro durabilities and/or in vivo persistence, and three out of the four types of carbon nanotubes tested (single-walled (CNTSW) and multi-walled (CNTTANG2, CNTSPIN)) showed no, or minimal, loss of mass or change in fibre length or morphology when examined by electron microscopy. However, the fourth type [multi-walled (CNTLONG1)] lost 30% of its original mass within the first three weeks of incubation, after which there was no further loss. Electron microscopy of CNTLONG1 samples incubated for 10 weeks confirmed that the proportion of long fibres had decreased compared to samples briefly exposed to the Gambles solution. This loss of mass and fibre shortening was accompanied by a loss of pathogenicity when injected into the peritoneal cavities of C57Bl/6 mice compared to fibres incubated briefly. CNTSW did not elicit an inflammogenic effect in the peritoneal cavity assay used here. CONCLUSIONS: These results support the view that carbon nanotubes are generally durable but may be subject to bio-modification in a sample-specific manner. They also suggest that pristine carbon nanotubes, either individually or in rope-like aggregates of sufficient length and aspect ratio, can induce asbestos-like responses in mice, but that the effect may be mitigated for certain types that are less durable in biological systems. Results indicate that durable carbon nanotubes that are either short or form tightly bundled aggregates with no isolated long fibres are less inflammogenic in fibre-specific assays.