EU → US Device Access

Which European device manufacturers are registered in the United States?

Type at least three letters. 4 871 European manufacturers with five or more devices declared.

Two public registers, no shared identifier, and nobody joins them. This site compares 4 871 European manufacturers from the EU medical device database against the U.S. FDA establishment register, and publishes what the comparison finds — including what it cannot settle.

Matched in the FDA register1 193 24.5 %
Not found3 448 70.8 %
Similar name, unverified212 4.4 %
⚠️ Three states, not two. The two registers share no common key: the comparison is made on company names. A name that does not match may mean the company is absent, or simply written differently. We keep those apart — a list that confused them was measured to be 10 % wrong.

The data behind these pages

The full comparison — every European manufacturer, its three-state FDA status, the matched FDA number where there is one, device counts and risk classes — is available as a file. So is the list of the 3 448 manufacturers not found in the U.S. register.

Pro access and prices → What the dataset contains →

By country

CountryManufacturers Not found in FDAShare
Germany1 08470465 %
Italy80865181 %
France49135071 %
United Kingdom40421252 %
Poland27926394 %
Spain25719174 %
Netherlands21516677 %
Switzerland1959147 %
Sweden1638653 %
Finland1319673 %
Denmark1137566 %
Belgium1067066 %
Czechia1008686 %
Austria815568 %
Ireland653452 %
Portugal585697 %
Hungary544889 %
Greece462759 %
Romania464496 %
Slovakia433888 %
Norway312374 %
Slovenia232191 %
Croatia181794 %
Lithuania171588 %
Bulgaria10880 %
Liechtenstein10220 %
Latvia8788 %
Estonia55100 %
Cyprus33100 %
Iceland3267 %
Luxembourg2150 %
Malta2150 %

What this is for

A manufacturer registered in Europe but absent from the U.S. register has not entered the American market — or has entered it under another name. For anyone selling regulatory, distribution or market-access services, that distinction is the difference between a prospect and a mistake.

The dataset, its limits and its price →