The regulatory ancestral network of surgical meshes

All surgical meshes entering the U.S. market have been cleared for clinical use by the 510(k) process of the Food and Drug Administration (FDA), in which devices simply require proof of "substantial equivalence" to predicate devices, without the need for clinical trials. However, recalled meshes associated with adverse effects may, indirectly, continue to serve as predicates for new devices raising concerns over the safety of the 510(k) route.

Here we assess the potential magnitude of this problem by determining the ancestral network of equivalence claims linking recently cleared surgical meshes. Using the FDA website we identified all surgical meshes cleared by the 510(k) route between January 2013 and December 2015 along with all listed predicates for these devices. Using a network approach, we trace the ancestry of predicates across multiple generations of equivalence claims and identify those meshes connected to devices that have since recalled from the market along with the reason for their recall.

We find that the 77 surgical meshes cleared between 2013 and 2015 are based on 771 interconnected predicate claims of equivalence from 400 other devices. The vast majority of these devices (97%) are descended from only six surgical meshes that were present on the market prior to 1976. One of these ancestral meshes alone, provided the basis of 183 subsequent devices. Furthermore, we show that 16% of recently cleared devices are connected through equivalence claims to the 3 predicate meshes that have been recalled for design and material related flaws causing serious adverse events. Taken together, our results show that surgical meshes are connected through a tangled web of equivalency claims and many meshes recently cleared by the FDA have connections through chains of equivalency to devices which have been recalled from the market due to concerns over clinical safety. These findings raise concerns over the efficacy of the 510(k) route in ensuring patient safety.