For regulatory and clinical affairs teams pursuing a De Novo classification, understanding what types of evidence FDA will accept is foundational to study design and submission strategy. As sponsors increasingly operate in global markets and accumulate post-market data from international registries, CE-marked product experience, and foreign adverse-event systems, the question of whether that evidence can carry weight in a US De Novo decision has material implications for development timelines and data-collection investments.
This analysis examines the FDA De Novo decision-summary corpus to identify submissions in which real-world evidence or real-world data collected outside the United States contributed to FDA's benefit-risk assessment and classification reasoning. It documents the confirmed cases, characterizing each by the registry or study involved, the relevant geography, sample size, data type, and the evidentiary role the data played in the agency's determination.
Want Rhizome's help on your own question? Try it for free.
When has the FDA accepted out-of-US real-world evidence for a De Novo?
Direct answer
Yes, the FDA has repeatedly accepted real-world evidence generated outside the United States as part of De Novo classification decisions, most often as supportive clinical evidence within the benefit-risk assessment rather than as the sole pivotal dataset. The clearest examples span 2019 to 2024 and include the Surfacer Inside-Out Access Catheter (DEN190038), DuraGraft (DEN230002), the Symani Surgical System (DEN230032), the Comaneci Embolization Assist Device (DEN170064), and the Apollo ESG/REVISE systems (DEN210045). A recurring pattern: FDA accepted the ex-US real-world data to support safety and effectiveness, flagged limited generalizability to the US population, and then imposed a US postmarket study as a special control.
Confirmed cases
Comaneci Embolization Assist Device (DEN170064, decision April 24, 2019)
This is the earliest of the confirmed cases. FDA stated that "clinical data from an outside the United States (US) post-market retrospective collection of intracranial aneurysms treated with the subject device was used to support the safety and effectiveness" of the device 4. The data came from a retrospective post-market case series at two ex-US sites, the Walton Centre in Liverpool, United Kingdom and University Hospital St. Ivan Rilski in Sofia, Bulgaria, covering 63 consecutively treated patients with 64 intracranial aneurysms and follow-up to three months, collected on a pre-specified case report form 4. FDA relied on the reported technical success (93.65%, 59/63) and serious neurological adverse-event rate (11.1%, 7/63) in its benefit-risk determination, while explicitly cautioning that the neurological adverse-event profile "was obtained from only two sites, both of which are located outside the US," creating "additional uncertainty" about predicting US safety 4. Because of that limitation, FDA required a post-market, prospective, multicenter study in US patients to "supplement existing evidence" 4. The device was granted De Novo and classified Class II under 21 CFR 882.5955 4.
Surfacer Inside-Out Access Catheter System (DEN190038, decision February 10, 2020)
FDA considered the International Post Market SAVE Registry, described as a "prospective, single-arm, post market, multicenter, international registry" reflecting routine clinical practice outside the United States 1. The registry enrolled 30 subjects at five centers in Germany, Italy, Austria, and Paraguay 1. FDA characterized it as "additional clinical data" and concluded that "the results obtained for subjects enrolled in the SAVE Registry support the safety and efficacy of the Surfacer Inside-Out procedure" 1. Because the US pivotal IDE cohort was small (30 subjects) and heavily proctored, FDA required a US postmarket evaluation to "obtain a more comprehensive safety profile of the device by collecting short-term safety data associated with real-world US use of the device in the absence of proctors" 1.
Apollo ESG / ESG Sx / REVISE / REVISE Sx Systems (DEN210045, decision July 12, 2022)
FDA stated plainly that "real-world evidence used to support a reasonable assurance of safety and effectiveness included real-world registry data and published clinical literature" 5. One registry included 295 subjects enrolled at a private bariatric practice "in the US and in Brazil," making this a mixed US/ex-US real-world dataset 5. FDA weighted the evidence cautiously, noting that "registry data from OUS locations (i.e., Brazil) may not be representative of outcomes in the US patient population" 5. The real-world registry data functioned as supporting evidence alongside the pivotal MERIT trial, literature, and post-market surveillance in the overall benefit-risk determination that supported Class II classification with special controls 5.
DuraGraft (DEN230002, decision October 4, 2023)
This case features two distinct ex-US real-world sources. The primary source was the DuraGraft EU Registry, "an ongoing all-comers European prospective registry" used as a post-approval study for the CE mark, enrolling 2,964 patients at 45 centers across eight countries: Austria, Germany, Ireland, Italy, Spain, Switzerland, Turkey, and the United Kingdom, with one-year follow-up continuing to five years 2. FDA said "the device safety is supported by data collected in the clinical studies described above" and compared registry MACE and mortality (30-day death rate 2.3%, one-year mortality 4.5%) against contemporary CABG literature and a propensity-matched US STS registry comparator 2. A second ex-US source was a prospective, multicenter randomized controlled trial (Perrault 2019) at seven sites in Canada, Ireland, and Denmark, enrolling 125 patients 2. Both ex-US datasets supported the safety evidence and the clinical-data special control underpinning Class II classification, with FDA concluding "the probable benefits outweigh the probable risks" 2.
Symani Surgical System (DEN230032, decision April 5, 2024)
FDA relied on the "Symani study," described as an "OUS, prospective, single arm, multicenter, post-market clinical follow-up (PMCF) study" conducted in Europe in a "real-world setting," with retrospective data collection added by protocol addendum 3. The analysis included 93 evaluable patients for free-flap surgery of the breast and extremities and 70 evaluable patients for lymphatic surgery 3. FDA stated "the Symani study provides evidence of the safety and effectiveness of the Symani Surgical System" 3. The submission also referenced supplementary real-world commercial procedure data from six European centers (77 free-flap and 22 lymphatic procedures) 3. FDA noted the ex-US study population differed demographically from the US population and therefore required a US postmarket study evaluating anastomosis-specific reoperation rate in a representative US population 3.
Patterns worth noting for a De Novo strategy
- Ex-US RWE has been accepted as supporting, not stand-alone, evidence. In every confirmed case the ex-US data sat alongside a pivotal study, literature, or nonclinical testing rather than serving as the sole basis for the decision 12345.
- Post-market registries and retrospective post-market case series are the dominant data types, with CE-mark post-approval registries (DuraGraft, Symani) and international post-market registries (Surfacer) recurring 123.
- Generalizability to the US population is FDA's central concern. Reviewers repeatedly flagged that ex-US data may not represent US practice or demographics (Comaneci, Apollo, Symani) 345.
- A US postmarket study is the common mitigation. For Comaneci, Surfacer, and Symani, FDA converted the generalizability gap into a required US postmarket study as a condition tied to the classification 134.
Limitations
This review reflects De Novo decision summaries retrievable in the dataset and the terminology those summaries use; a submission that used ex-US real-world data without describing it in RWE/RWD language could be missed. Sample sizes, dates, and geographies above are taken directly from each decision summary. A reader who needs an exhaustive census, or who wants the full evidence tables and special-control text for any single device, can ask Rhizome to pull that decision summary in detail.