How to find a predicate device for a GEI 510(k)

Product code GEI covers electrosurgical and radiofrequency devices, from laparoscopic instruments to RF skin systems. There are 2,316 GEI clearances in the FDA 510(k) index we track. This page shows you how to work that pool, with real recent clearances below.

What the GEI predicate pool looks like

Over the trailing 24 months to September 2026, FDA decided 106 GEI submissions: median review 135 days, 90th percentile 267 days. The most frequent clearers in the code are Arthrocare (64 clearances on file), Valleylab (60), and Conmed (43). Plan a longer predicate runway here than in faster codes: the median GEI review runs about six weeks longer than GEX.

Recent GEI clearances to investigate as predicates

Earlier clearances in the same product code are candidates to investigate, not predicate determinations. Verify any predicate claim against the official 510(k) summary.

• K254185, cleared 2026-09-11: Potenza Prime, Jeisys Medical Incorporated.
• K262364, cleared 2026-09-02: Ellian Enigma RF Device, Ellian Corp.
• K253859, cleared 2026-08-27: Scarlet Pro (Scarlet Pro), VIOL Co., Ltd.
• K252535, cleared 2026-08-25: Inno-Switch Laparoscopic Instruments, Taiwan Surgical Corporation.
• K261179, cleared 2026-07-22: AlphaTriPart Monopolar MIC Instruments, Gimmi GmbH.

How to run the search

1. Enter your closest known 510(k) number in the free medical device predicate search. It pulls the FDA record and lists earlier same-code clearances as candidates.
2. Open each candidate's FDA record and read its 510(k) summary for the listed predicate. Only the summary establishes the relationship.
3. Watch the code afterward: see check any product code for the latest clearances, or read how competitive intelligence works.

Comparing codes? See the GEX predicate guide and the general how to find a predicate device for a 510(k) walkthrough.

Data: U.S. FDA via openFDA (public domain), plus our 510(k) index of 176,140 records and review benchmarks computed September 2026. Not regulatory advice. Assume all results are unvalidated.