B. Braun Melsungen AG — Patent Strategy Analysis 2014–2024
A DOCDB family-level analysis of the B. Braun group portfolio — 1,450 patent families filed at 36 patent offices, grouped into eight filing entities including Aesculap AG, with 93 % of the portfolio in medical technology.
Executive Summary
B. Braun group patent activity, DOCDB family level, first filing year 2014–2024
Between 2014 and 2024 the B. Braun group filed 1,450 DOCDB patent families comprising 5,151 patent applications at 36 patent offices. The scope of this report is the corporate group, not the parent company alone: eight filing entities are consolidated, among them Aesculap AG, which at 606 families is the group’s largest single filer — ahead of the Melsungen headquarters with 437 families and B. Braun Avitum AG with 247.
Filing volume rose over the decade. In 2014–2018 the group first-filed an average of 111.6 families per year; in 2019–2022 the average was 142.0 families per year, an increase of 27 %. The two most recent years, 2023 and 2024, are shown throughout this report but are not used for trend claims: applications published later than the PATSTAT Spring 2026 cut-off are not yet in the data.
The portfolio is narrow by technology and wide by geography. Measured on WIPO technology fields, 1,355 of 1,450 families (93.4 %) fall into Medical technology — yet all 35 WIPO fields appear somewhere in the portfolio, from Computer technology (104 families) to Semiconductors (7). The two dominant IPC subclasses are A61B (Diagnosis; Surgery; Identification, 692 families) and A61M (Devices for introducing media into the body, 609 families).
Inside that medical core the balance shifted. Extracorporeal blood treatment — the dialysis business — declined: families per year in IPC A61M 1/16 (dialysis membranes) fell from 17.6 in 2014–2018 to 12.0 in 2019–2022. Surgical instrumentation, navigation and orthopaedics rose in the same window: IPC A61B 34/30 (surgical robots) grew from 1 family in 2014–2018 to 20 in 2019–2022, and healthcare informatics (G16H) went from 3.2 to 12.0 families per year.
The filing route changed as well. The share of families carrying a WO (PCT) member rose from 42.7 % in 2014–2018 to 63.2 % in 2019–2022, while direct national filings in Japan fell from 19.7 % to 6.2 % of period families and in Spain from 22.2 % to 12.7 %. Part of that decline is a genuine route change and part of it is timing — national-phase entries from recent PCT filings are due up to 30 months after priority and are therefore under-represented in the current edition.
External collaboration is rare. Only 41 of 1,450 families (2.8 %) carry a co-applicant from outside the group, spread over 57 distinct organisations; the largest single partner, Universitat Politècnica de Catalunya, appears on 14 families. A further 16 families are co-filed between two B. Braun group entities and are treated here as internal, not as collaboration.
The decade reads as a shift of centre of gravity inside one company: away from the dialysis and infusion technology that defined B. Braun’s classic portfolio, toward surgical instrumentation, navigation, orthopaedic implants and device software — the Aesculap side of the group. Aesculap alone accounts for 606 of 1,450 families, and the German inventor base reflects the same weight: on geocoded EP applications, Baden-Württemberg (Tuttlingen) is linked to 426 families and Hessen (Melsungen) to 325.
Corporate Entities and Scope
Which legal entities the 1,450 families were filed by — and how the group boundary was drawn
B. Braun is a family-owned group that files under many names. This report consolidates the group: B. Braun Melsungen AG as parent, plus the operating subsidiaries that file under a B. Braun name, plus Aesculap AG, the wholly owned surgical-technology arm in Tuttlingen. Aesculap files under its own brand and shares no name element with the parent, so it is invisible to any search anchored on “Braun” — and it is the largest filer in the group.
Aesculap AG accounts for 606 of 1,450 families (41.8 %), more than the Melsungen headquarters (437, 30.1 %). B. Braun Avitum AG, the dialysis division, contributes 247 families. B. Braun New Ventures GmbH — the group’s innovation unit — reaches 59 families.
Entity totals exceed the portfolio total. The eight entity rows sum to 1,466 families against 1,450 distinct families. The difference is exactly the 16 families co-filed by two group entities, which appear under both. Percentages in this section are calculated against the 1,450 distinct families.
Data table: Patent families and applications by group entity
| Filing entity (consolidated name variants) | Families | Applications | Share of 1,450 |
|---|---|---|---|
| Aesculap AG (surgical technology, Tuttlingen) | 606 | 2,254 | 41.8 % |
| B. Braun Melsungen AG (headquarters) | 437 | 1,552 | 30.1 % |
| B. Braun Avitum AG (dialysis) | 247 | 834 | 17.0 % |
| B. Braun New Ventures GmbH | 59 | 160 | 4.1 % |
| B. Braun Medical (US / FR / other) | 54 | 188 | 3.7 % |
| B. Braun China entities (Suzhou, Jiangsu, Beijing, Shanghai) | 38 | 38 | 2.6 % |
| B. Braun Surgical S.A. (Spain) | 21 | 111 | 1.4 % |
| B. Braun Miethke GmbH | 4 | 24 | 0.3 % |
| Distinct families (group total) | 1,450 | 5,151 | 100 % |
How the group boundary was drawn
The scope was built on person_name in tls206_person, not on the
harmonised han_name, and then cross-checked against han_name in
both directions. A discovery pass over every applicant record containing
Braun or Aesculap, plus Chinese, Japanese and Korean renderings, returned
145 applicant name records with at least one family in the period. These were reviewed one by one.
Included are all spellings of B. Braun and Aesculap entities, including office-side renderings that no Latin-script search reaches — Chinese 贝朗 (B. Braun) and 蛇牌 (Aesculap), and Japanese katakana forms such as ベー・ブラウン・メルズンゲン and ビー．ブラウン アビタム. Excluded are unrelated companies that share the name stem: Braun GmbH / Braun AG and De’Longhi Braun Household (personal-care appliances, Kronberg), Hartmann & Braun (instrumentation), The Braun Corporation (US mobility equipment), Braun Elevator, TricorBraun, Asea Braun Boveri, the Braunkohle and Braunschweig name families, individual inventors surnamed Braun, and Aesculapius Farmaceutici / Aesculapius Limited, which are not part of the group.
What the han_name cross-check showed
-
No contamination. The harmonised groups used by the accepted variants
contain no foreign organisation: a scope built on
han_namereturns 0 families that theperson_namescope does not also contain. -
Near-complete, but not complete. The
han_nameroute returns 1,447 families against 1,450 — it misses 3 families whose only applicant records are non-Latin renderings that carry themselves as their ownhan_name. -
The real han_name trap here is grouping, not spelling. Aesculap records
harmonise to
AESCULAP AG, which contains no “Braun” string at all. A harmonised-name search anchored on the parent company alone therefore does not reach any of the 606 Aesculap families — 41.8 % of the portfolio.
Filing Trend 2014–2024
Patent families by first filing year of the family, all group entities combined
Activity is higher in the second half of the decade than in the first. Over the complete years, the group first-filed 558 families in 2014–2018 (111.6 per year) and 568 families in 2019–2022 (142.0 per year) — an increase of 27.2 % in annual average. The trough is 2017 with 94 families; the highest complete year is 2021 with 165.
Application volume moves with family volume but not proportionally: 5,151 applications across 1,450 families, an average of 3.55 own applications per family. The peak application year is 2021 with 604 applications.
* 2023 and 2024 are provisional. A patent application is normally published 18 months after its priority date, so families first filed in these years are still entering the database. The 2024 value in particular is a lower bound and is excluded from all trend statements in this report (PATSTAT Global, Spring 2026).
Data table: Filing trend 2014–2024
| First filing year | Patent Families | Year on year | Own applications filed |
|---|---|---|---|
| 2014 | 129 | — | 431 |
| 2015 | 111 | −14.0 % | 474 |
| 2016 | 119 | +7.2 % | 514 |
| 2017 | 94 | −21.0 % | 389 |
| 2018 | 105 | +11.7 % | 425 |
| 2019 | 125 | +19.0 % | 480 |
| 2020 | 134 | +7.2 % | 549 |
| 2021 | 165 | +23.1 % | 604 |
| 2022 | 144 | −12.7 % | 528 |
| 2023* | 193 | +34.0 % | 483 |
| 2024* | 131 | −32.1 % | 274 |
| Total | 1,450 | — | 5,151 |
A family is counted once, in the year of its earliest filing anywhere in the world. That earliest year is taken over all members of the family, including members filed before 2014; families whose earliest filing predates 2014 are outside this report even if they were extended during the period. The yearly figures therefore add up exactly to the portfolio total.
Geographic Filing Strategy
Where the 1,450 families are filed, and how the office mix shifted between 2014–2018 and 2019–2022
Germany is the most used office in absolute terms: 1,085 of 1,450 families (74.8 %) have a DE filing, ahead of EP (996, 68.7 %) and US (849, 58.6 %). The clearest movement between the two complete periods is the international route: the share of families with a WO (PCT) application rose from 42.7 % to 63.2 %, and China from 30.8 % to 39.8 %.
Direct national filings outside the core moved the other way — Japan from 19.7 % to 6.2 % of period families, Spain from 22.2 % to 12.7 %, Singapore from 3.9 % to 0.7 %, and Russia from 4.3 % to none. This is consistent with a substitution of direct national filings by the PCT route, but it is not proof of one: see the caveat below.
National-phase timing distorts the second period. A PCT application enters the national phase up to 30 months after priority, and the resulting national application is published later still. For families first filed in 2021 or 2022, many national-phase filings are therefore not yet in the Spring 2026 edition. The declining national shares in the 2019–2022 column are a lower bound; the rising WO share is not affected by this. Both period boundaries were placed on complete filing years — 2023 and 2024 are deliberately outside this comparison.
Data table: Top filing offices 2014–2024
| Office | Families | Applications | Share of 1,450 families |
|---|---|---|---|
| DE — Germany | 1,085 | 1,085 | 74.8 % |
| EP — European Patent Office | 996 | 1,070 | 68.7 % |
| US — United States | 849 | 908 | 58.6 % |
| WO — PCT (WIPO) | 731 | 746 | 50.4 % |
| CN — China | 465 | 483 | 32.1 % |
| ES — Spain | 197 | 213 | 13.6 % |
| JP — Japan | 167 | 183 | 11.5 % |
| AU — Australia | 71 | 91 | 4.9 % |
| PL — Poland | 56 | 58 | 3.9 % |
| BR — Brazil | 51 | 53 | 3.5 % |
| KR — Korea | 36 | 38 | 2.5 % |
| CA — Canada | 29 | 32 | 2.0 % |
| SG — Singapore | 26 | 32 | 1.8 % |
| RU — Russia | 24 | 25 | 1.7 % |
| PT — Portugal | 20 | 22 | 1.4 % |
| MX — Mexico | 19 | 20 | 1.3 % |
| DK — Denmark | 18 | 20 | 1.2 % |
| 19 further offices (HK, HU, IL, MY, ZA, FR, GB, TW, EA, TR, IT, CL, AT, DO, FI, NL, BE, SA, SE) | ≤ 12 each | — | — |
Family counts per office add up to more than 1,450 because one family is normally filed at several offices. 36 distinct offices appear in the portfolio. The WO row counts the 731 families in which a group entity is itself the PCT applicant; the PCT rate of 50.8 % quoted in the key figures counts the 737 families that contain any WO member.
Data table: Office coverage by period (share of period families)
| Office | 2014–2018 families | Share of 558 | 2019–2022 families | Share of 568 | Change |
|---|---|---|---|---|---|
| DE | 431 | 77.2 % | 435 | 76.6 % | −0.6 pp |
| EP | 383 | 68.6 % | 433 | 76.2 % | +7.6 pp |
| US | 363 | 65.1 % | 372 | 65.5 % | +0.4 pp |
| WO | 238 | 42.7 % | 359 | 63.2 % | +20.5 pp |
| CN | 172 | 30.8 % | 226 | 39.8 % | +9.0 pp |
| ES | 124 | 22.2 % | 72 | 12.7 % | −9.5 pp |
| JP | 110 | 19.7 % | 35 | 6.2 % | −13.5 pp |
| AU | 27 | 4.8 % | 38 | 6.7 % | +1.9 pp |
| PL | 32 | 5.7 % | 23 | 4.0 % | −1.7 pp |
| BR | 21 | 3.8 % | 30 | 5.3 % | +1.5 pp |
| KR | 11 | 2.0 % | 21 | 3.7 % | +1.7 pp |
| SG | 22 | 3.9 % | 4 | 0.7 % | −3.2 pp |
| RU | 24 | 4.3 % | 0 | 0.0 % | −4.3 pp |
Technology Fields (WIPO 35)
The portfolio mapped onto the 35 WIPO technology fields, aggregated over all family members
The portfolio is dominated by a single field: 1,355 of 1,450 families (93.4 %) carry the WIPO field Medical technology. Everything else is small by comparison — the second field, Handling, reaches 135 families (9.3 %).
At the same time all 35 WIPO fields are represented, down to Basic communication processes with a single family. The fastest growing field in relative terms is Computer technology: 5.4 families per year in 2014–2018 against 12.5 in 2019–2022. Electrical machinery, apparatus, energy moves from 2.0 to 7.3 families per year. Both are small bases, so these are directional signals rather than established trends.
Non-exclusive assignment. One patent family can be assigned to several WIPO fields, because the assignment derives from all IPC symbols on all family members. The column below therefore sums to more than 1,450. For the same reason no pie or doughnut chart is used here.
Data table: All 35 WIPO technology fields
| Sector | WIPO field | Families | 2014–2018 (per year) | 2019–2022 (per year) |
|---|---|---|---|---|
| Instruments | Medical technology | 1,355 | 525 (105.0) | 536 (134.0) |
| Mechanical engineering | Handling | 135 | 58 (11.6) | 45 (11.3) |
| Instruments | Measurement | 112 | 42 (8.4) | 45 (11.3) |
| Electrical engineering | Computer technology | 104 | 27 (5.4) | 50 (12.5) |
| Mechanical engineering | Mechanical elements | 96 | 36 (7.2) | 41 (10.3) |
| Chemistry | Chemical engineering | 88 | 40 (8.0) | 26 (6.5) |
| Mechanical engineering | Machine tools | 53 | 21 (4.2) | 25 (6.3) |
| Mechanical engineering | Engines, pumps, turbines | 47 | 19 (3.8) | 16 (4.0) |
| Electrical engineering | Electrical machinery, apparatus, energy | 46 | 10 (2.0) | 29 (7.3) |
| Mechanical engineering | Other special machines | 46 | 19 (3.8) | 17 (4.3) |
| Chemistry | Pharmaceuticals | 33 | 12 (2.4) | 11 (2.8) |
| Instruments | Control | 32 | 10 (2.0) | 14 (3.5) |
| Electrical engineering | Audio-visual technology | 30 | 14 (2.8) | 13 (3.3) |
| Instruments | Analysis of biological materials | 28 | 13 (2.6) | 12 (3.0) |
| Chemistry | Environmental technology | 23 | 7 (1.4) | 5 (1.3) |
| Electrical engineering | Telecommunications | 22 | 9 (1.8) | 10 (2.5) |
| Electrical engineering | Digital communication | 22 | 3 (0.6) | 11 (2.8) |
| Chemistry | Surface technology, coating | 21 | 9 (1.8) | 6 (1.5) |
| Instruments | Optics | 21 | 8 (1.6) | 10 (2.5) |
| Chemistry | Materials, metallurgy | 21 | 7 (1.4) | 11 (2.8) |
| Chemistry | Basic materials chemistry | 15 | 6 (1.2) | 7 (1.8) |
| Chemistry | Biotechnology | 10 | 3 (0.6) | 4 (1.0) |
| Electrical engineering | IT methods for management | 10 | 1 (0.2) | 7 (1.8) |
| Chemistry | Organic fine chemistry | 10 | 1 (0.2) | 7 (1.8) |
| Other fields | Civil engineering | 10 | 5 (1.0) | 4 (1.0) |
| Other fields | Furniture, games | 9 | 4 (0.8) | 5 (1.3) |
| Electrical engineering | Semiconductors | 7 | 3 (0.6) | 4 (1.0) |
| Chemistry | Macromolecular chemistry, polymers | 6 | 2 (0.4) | 4 (1.0) |
| Mechanical engineering | Transport | 6 | 1 (0.2) | 3 (0.8) |
| Mechanical engineering | Textile and paper machines | 6 | 3 (0.6) | 2 (0.5) |
| Other fields | Other consumer goods | 5 | 2 (0.4) | 3 (0.8) |
| Mechanical engineering | Thermal processes and apparatus | 4 | 2 (0.4) | 0 (0.0) |
| Chemistry | Food chemistry | 3 | 1 (0.2) | 2 (0.5) |
| Chemistry | Micro-structural and nano-technology | 2 | 0 (0.0) | 1 (0.3) |
| Electrical engineering | Basic communication processes | 1 | 0 (0.0) | 1 (0.3) |
The two period columns cover 5 and 4 complete years respectively; the value in brackets is families per year, which is the comparable figure. 2023 and 2024 are not part of either period.
Technology Classes (IPC)
IPC subclasses and main groups, aggregated over every IPC symbol on every member of the 1,450 families
Two subclasses carry the portfolio: A61B (Diagnosis; Surgery; Identification) with 692 families (47.7 %) and A61M (Devices for introducing media into or onto the body) with 609 families (42.0 %). Between them they cover the surgical and the infusion/dialysis halves of the group.
The clearest structural movement sits outside the classic medical subclasses: G16H — Healthcare informatics rises from 3.2 families per year in 2014–2018 to 12.0 in 2019–2022, and G06K — graphical data reading, identification from 1.2 to 5.5. In the same window the separation subclass B01D — filters and membranes, the classic dialysis technology — falls from 5.4 to 3.0 families per year.
Multi-assignment. A family normally carries several IPC symbols, and this analysis takes every symbol from every family member (offices classify the same family differently). Family counts per class therefore sum well above 1,450 and cannot be read as shares of a whole. Horizontal bars are used deliberately; a pie or doughnut chart would be wrong for this data.
Data table: IPC subclasses with 15 or more families
| IPC subclass | Title | Families | 2014–2018 | 2019–2022 |
|---|---|---|---|---|
| A61B | Diagnosis; Surgery; Identification | 692 | 257 | 308 |
| A61M | Devices for introducing media into, or onto, the body | 609 | 252 | 208 |
| A61L | Sterilising; Materials for prostheses; Disinfection | 183 | 92 | 66 |
| A61F | Filters implantable into blood vessels; Prostheses | 176 | 70 | 75 |
| G16H | Healthcare informatics | 103 | 16 | 48 |
| A61J | Containers for medical or pharmaceutical purposes | 101 | 35 | 36 |
| B65D | Containers for storage or transport of articles | 96 | 48 | 32 |
| B01D | Separation | 59 | 27 | 12 |
| G01N | Investigating or analysing materials | 57 | 25 | 19 |
| G06F | Electric digital data processing | 51 | 20 | 17 |
| F04B | Positive-displacement machines for liquids; Pumps | 40 | 17 | 12 |
| B29C | Shaping or joining of plastics | 32 | 16 | 11 |
| F16L | Pipes; Joints or fittings for pipes | 32 | 10 | 17 |
| A61K | Preparations for medical, dental or toiletry purposes | 31 | 12 | 9 |
| G06K | Graphical data reading; Presentation of data | 31 | 6 | 22 |
| F16K | Valves; Taps; Cocks | 27 | 18 | 8 |
| F16B | Devices for fastening or securing constructional elements | 26 | 6 | 12 |
| A61C | Dentistry; Oral or dental hygiene | 25 | 5 | 16 |
| B25J | Manipulators; Chambers with manipulation devices | 24 | 2 | 9 |
| B65B | Machines or methods for packaging articles | 22 | 11 | 4 |
| A61P | Specific therapeutic activity of chemical compounds | 21 | 7 | 7 |
| G06T | Image data processing or generation | 20 | 3 | 9 |
| B01J | Chemical or physical processes, e.g. catalysis | 17 | 7 | 9 |
| B33Y | Additive manufacturing | 17 | 4 | 8 |
| G01M | Testing static or dynamic balance of machines | 17 | 4 | 5 |
| H01R | Electrically-conductive connections | 16 | 2 | 9 |
| G02B | Optical elements, systems or apparatus | 16 | 7 | 7 |
| F04C | Rotary-piston positive-displacement machines | 16 | 5 | 4 |
| G01L | Measuring force, stress, torque, work | 16 | 3 | 11 |
| H04L | Transmission of digital information | 15 | 1 | 6 |
Largest IPC main groups
One level deeper, the main groups show what the two dominant subclasses actually contain: surgical instruments and accessories on the A61B side, dialysis and infusion technology on the A61M side.
Data table: 25 largest IPC main groups
| IPC main group | Title | Families |
|---|---|---|
| A61B 17/00 | Surgical instruments, devices or methods | 204 |
| A61M 1/16 | Blood treatment with membranes | 181 |
| A61M 1/14 | Dialysis systems; Artificial kidneys; Blood oxygenators | 132 |
| A61B 90/00 | Instruments or accessories for surgery or diagnosis | 130 |
| A61M 5/142 | Pressure infusion, e.g. using pumps | 110 |
| A61M 1/36 | Treatment of blood in a by-pass of the natural circulation | 109 |
| A61M 5/14 | Infusion devices; Blood infusion | 102 |
| A61M 25/00 | Catheters; Hollow probes | 97 |
| A61L 2/26 | Sterilising accessories | 89 |
| A61B 50/30 | Containers for packaging or sterilising surgical instruments | 89 |
| A61F 2/46 | Special tools for implanting artificial joints | 86 |
| A61F 2/30 | Joint prostheses | 84 |
| A61B 17/16 | Instruments for osteoclasis; Drills | 81 |
| A61B 34/20 | Surgical navigation systems; Tracking devices | 81 |
| A61M 25/06 | Body-piercing guide needles | 79 |
| A61B 17/56 | Surgical instruments for treatment of bones | 78 |
| A61M 39/10 | Tube connectors or tube couplings | 77 |
| A61M 1/00 | Suction or pumping devices for medical purposes | 75 |
| A61M 5/168 | Means for controlling media flow to the body | 73 |
| A61F 2/38 | Prostheses for elbows or knees | 69 |
| A61B 17/70 | Spinal positioners or stabilisers | 67 |
| A61B 50/00 | Containers, covers or holders for medical instruments | 58 |
| A61B 17/28 | Surgical forceps | 52 |
| A61M 5/32 | Needles; Details of needles | 47 |
| A61B 34/30 | Surgical robots | 47 |
Grant Rates by Office
Applications of the 2014–2020 filing cohort and the share of them recorded as granted
At the offices that examine substantively, the group converts filings into grants at a high rate: EP 72.7 % (459 of 631 applications), US 75.1 % (443 of 590) and JP 79.5 % (105 of 132). China is markedly lower at 46.1 % (131 of 284).
The German figure stands apart: 38 of 640 DE applications (5.9 %) are recorded as granted. The DE filing is the most frequent filing in the portfolio (1,085 families) and the least frequently granted. This pattern is consistent with the German application being used as a priority filing whose subject matter is then pursued through EP or PCT rather than to a German national patent — but the data here shows the outcome, not the intention.
Across the whole period, 747 of 1,450 families (51.5 %) contain at least one granted member and 1,988 of 5,151 applications (38.6 %) are recorded as granted.
Pendency and coverage caveats. Grant rates are restricted to the
2014–2020 filing cohort; applications filed from 2021 onward are still in examination and
would depress every figure. Offices that register rather than examine (ES until the 2017
reform, PL, PT, DK, RU in this cohort) show rates at or near 100 % and are not comparable
with examining offices. Singapore shows 0 granted of 32 applications, which reflects
missing grant information in the source data rather than 32 refusals. Grant status is
taken from the granted flag on the application record; WO applications are
excluded because the PCT does not grant.
Data table: Grant rate by office, 2014–2020 filing cohort
| Office | Applications | Granted | Grant rate | Examination |
|---|---|---|---|---|
| DE — Germany | 640 | 38 | 5.9 % | substantive on request |
| EP — European Patent Office | 631 | 459 | 72.7 % | substantive |
| US — United States | 590 | 443 | 75.1 % | substantive |
| CN — China | 284 | 131 | 46.1 % | substantive |
| ES — Spain | 180 | 180 | 100.0 % | registration / partial |
| JP — Japan | 132 | 105 | 79.5 % | substantive |
| AU — Australia | 67 | 47 | 70.1 % | substantive |
| PL — Poland | 46 | 46 | 100.0 % | registration / partial |
| BR — Brazil | 43 | 22 | 51.2 % | substantive |
| SG — Singapore | 32 | 0 | n/a | grant data not covered |
| RU — Russia | 25 | 25 | 100.0 % | registration / partial |
| KR — Korea | 23 | 13 | 56.5 % | substantive |
| DK — Denmark | 20 | 20 | 100.0 % | registration / partial |
| PT — Portugal | 20 | 20 | 100.0 % | registration / partial |
| CA — Canada | 20 | 9 | 45.0 % | substantive |
Offices with fewer than 20 applications in the cohort are omitted. The “Examination” column is a reading aid for interpreting the rate, not a value taken from PATSTAT.
Collaboration Network
Co-applicants from outside the group — and the internal co-filings that are deliberately not counted as collaboration
B. Braun files almost entirely on its own account.
62 of 1,450 families (4.3 %) carry more than one applicant on
the application record. Measured independently,
41 families (2.8 % of the portfolio) have at least one applicant
outside the group scope, spread over 57 distinct co-applicant names after
normalisation, and 16 families are co-filed by two B. Braun group
entities. The three counts come from different fields — the
nb_applicants flag on the application, the applicant records themselves, and
the entity grouping — and are not additive.
Because this report treats the group as one applicant, those 16 families are counted as internal co-filings, not as collaboration, and every group spelling is excluded from the partner ranking below. Without that exclusion the group’s own subsidiaries would occupy the top of its own partner list.
The strongest external link is academic and Catalan: Universitat Politècnica de Catalunya appears on 14 families, more than three times any other partner. Together with Fundació Institut de Ciències Fotòniques (4 families) and ICREA (3), the Barcelona research cluster accounts for the majority of the group’s named academic co-filings; the group’s Spanish entity B. Braun Surgical S.A. accounts for 21 families in the portfolio. Industrial co-applicants are few and small: Koninklijke Philips N.V. and Nanjing Nuoyuan Medical Devices with 3 families each.
Data table: External co-applicants with 2 or more shared families
| Co-applicant | Country | Type | Shared families |
|---|---|---|---|
| Universitat Politècnica de Catalunya | ES | University | 14 |
| Fundació Institut de Ciències Fotòniques (ICFO) | ES | Research | 4 |
| Koninklijke Philips N.V. | NL | Company | 3 |
| Nanjing Nuoyuan Medical Devices Co., Ltd. | CN | Company | 3 |
| Institució Catalana de Recerca i Estudis Avançats (ICREA) | ES | Research | 3 |
| Eberhard Karls Universität Tübingen | DE | University | 2 |
| IHI Ionbond AG | CH | Company | 2 |
| Brunetti, Bruce | US | Individual | 2 |
| Co-applicant records in non-Latin script (not resolved) | — | — | 5 |
Name variants were merged by normalising case, accents and punctuation — the Barcelona university, for example, appears under five spellings in the raw data. The last row collects co-applicant records whose names contain no Latin characters and therefore cannot be merged by this normalisation; they are included in the totals of 41 families and 57 organisations but are not resolved to a single organisation name here. 48 further organisations appear on exactly one family each.
Self-inclusion check. The partner list is built by taking all further applicants on the group’s own applications and removing every person record inside the group scope, plus any remaining name containing Braun, Aesculap, 贝朗, 蛇牌 or ブラウン. No B. Braun or Aesculap entity appears in the table above.
Family Size and International Reach
How many patent applications each of the 1,450 families contains, counting every member worldwide
The 1,450 families contain 5,851 patent applications in total — including members filed by other applicants, against the 5,151 filed by the group itself — an average of 4.04 members per family; the median family has 4 members. The distribution has two peaks: 288 families (19.9 %) never leave a single office — the largest single bucket — while a second concentration sits at five and six members, and 589 families (40.6 %) reach five or more applications.
Seven families reach 15 or more members. Family size is a rough proxy for how much a company is willing to spend on an invention, and a fifth of the portfolio staying at one filing is consistent with the low German grant rate seen in the previous section: a share of the DE priority filings is never taken further.
Data table: Family size distribution
| Applications in the family | Families | Share | Cumulative share |
|---|---|---|---|
| 1 | 288 | 19.9 % | 19.9 % |
| 2 | 228 | 15.7 % | 35.6 % |
| 3 | 161 | 11.1 % | 46.7 % |
| 4 | 184 | 12.7 % | 59.4 % |
| 5 | 200 | 13.8 % | 73.2 % |
| 6 | 190 | 13.1 % | 86.3 % |
| 7 | 102 | 7.0 % | 93.3 % |
| 8 | 45 | 3.1 % | 96.4 % |
| 9 | 17 | 1.2 % | 97.6 % |
| 10 | 9 | 0.6 % | 98.2 % |
| 11 | 7 | 0.5 % | 98.7 % |
| 12 | 4 | 0.3 % | 99.0 % |
| 13 | 3 | 0.2 % | 99.2 % |
| 14 | 5 | 0.3 % | 99.5 % |
| 15 or more | 7 | 0.5 % | 100.0 % |
| All families | 1,450 | 100 % | — |
Family size counts every application in the DOCDB family, including members filed by other applicants and members filed outside the 2014–2024 window. Total members: 5,851.
Where the Portfolio Is Moving
IPC main groups compared between 2014–2018 and 2019–2022, normalised to families per year
The two halves of the decade describe a different company. Growing fastest, in families per year: A61B 34/30 surgical robots, from 1 family in the whole of 2014–2018 to 20 families in 2019–2022; A61B 17/16 bone drills and osteoclasis instruments, 3.2 → 13.3; A61F 2/38 knee and elbow prostheses, 2.0 → 10.0; A61M 5/142 pressure infusion pumps, 4.0 → 12.3.
Shrinking in the same window are the classic extracorporeal blood-treatment groups: A61M 1/16 dialysis membranes, 17.6 → 12.0 families per year (−32 %), A61M 1/14 dialysis systems 12.2 → 8.8 (−28 %), A61M 1/36 blood treatment in a by-pass 11.8 → 8.0 (−32 %), and A61M 25/06 body-piercing guide needles 9.6 → 6.0 (−38 %).
Read together with the subclass figures — G16H healthcare informatics 3.2 → 12.0 and B25J manipulators 0.4 → 2.3 families per year — the pattern is a shift of R&D weight from fluid handling toward instrumented, navigated and software-supported surgery. Groups with a base below 10 families in the first period are reported as absolute counts rather than as growth rates, because a percentage on such a base is not meaningful.
Data table: IPC main groups by period (families per year)
| IPC main group | Title | 2014–2018 | per year | 2019–2022 | per year |
|---|---|---|---|---|---|
| Rising | |||||
| A61B 34/30 | Surgical robots | 1 | 0.2 | 20 | 5.0 |
| A61B 17/16 | Instruments for osteoclasis; Drills | 16 | 3.2 | 53 | 13.3 |
| A61F 2/38 | Prostheses for elbows or knees | 10 | 2.0 | 40 | 10.0 |
| A61M 5/142 | Pressure infusion, e.g. using pumps | 20 | 4.0 | 49 | 12.3 |
| A61B 90/98 | Identification using electromagnetic means | 7 | 1.4 | 30 | 7.5 |
| A61F 2/46 | Special tools for implanting artificial joints | 24 | 4.8 | 41 | 10.3 |
| A61M 5/145 | Infusion using pressurised reservoirs | 8 | 1.6 | 26 | 6.5 |
| A61B 17/3201 | Surgical scissors | 11 | 2.2 | 25 | 6.3 |
| A61B 17/15 | Guides for surgical drills | 10 | 2.0 | 23 | 5.8 |
| A61B 90/90 | Identification means for patients or instruments | 11 | 2.2 | 21 | 5.3 |
| A61B 34/10 | Computer-aided planning, simulation, modelling | 5 | 1.0 | 17 | 4.3 |
| A61B 17/94 | Endoscopic surgical instruments | 3 | 0.6 | 14 | 3.5 |
| G16H 40/40 | Management of medical equipment or devices | 1 | 0.2 | 13 | 3.3 |
| Falling | |||||
| A61M 1/16 | Blood treatment with membranes | 88 | 17.6 | 48 | 12.0 |
| A61M 1/14 | Dialysis systems; Artificial kidneys | 61 | 12.2 | 35 | 8.8 |
| A61M 1/36 | Blood treatment in a by-pass of the circulation | 59 | 11.8 | 32 | 8.0 |
| A61M 25/06 | Body-piercing guide needles | 48 | 9.6 | 24 | 6.0 |
| A61B 17/70 | Spinal positioners or stabilisers | 37 | 7.4 | 17 | 4.3 |
| A61M 39/24 | Check- or non-return valves | 23 | 4.6 | 12 | 3.0 |
| A61B 17/88 | Means for implanting or extracting internal fixations | 22 | 4.4 | 7 | 1.8 |
| A61M 1/34 | Filtering material out of the blood | 22 | 4.4 | 14 | 3.5 |
| A61M 39/28 | Clamping means for flexible tubes | 14 | 2.8 | 6 | 1.5 |
| A61B 50/30 | Containers for packaging or sterilising instruments | 47 | 9.4 | 34 | 8.5 |
2014–2018 covers 5 complete years, 2019–2022 covers 4; only the per-year columns are directly comparable. 2023 and 2024 are excluded from this comparison because they are still incomplete. Families can carry several main groups, so the rows are not additive.
Where the Inventors Sit
Inventor countries across the portfolio, and German regions from the NUTS geocoding of EP applications
Invention is concentrated in Germany: 1,181 of 1,449 families with inventor data (81.5 %) name at least one German inventor. The second country is Malaysia with 70 families, ahead of the United States (67), Switzerland (61) and France (46).
Inside Germany the geocoded EP applications put Baden-Württemberg ahead of Hessen — 426 families against 325. Baden-Württemberg is the location of Aesculap in Tuttlingen, Hessen that of the Melsungen headquarters. The regional split mirrors the entity split of the portfolio: the surgical business generates more families than the parent company.
Coverage of the location data. Inventor country is present for most but not all records: 708 families contain at least one inventor record with no country code, so country counts are lower bounds. NUTS geocoding exists only for EP applications — 979 of the 1,070 EP applications in this portfolio (91.5 %) carry a NUTS code at level 3 or 4; German national applications have no geocoding at all. A family with inventors in several countries or regions is counted in each, so the columns sum above the family total.
Data table: Inventor countries
| Country | Families with an inventor there | Share of 1,449 |
|---|---|---|
| DE — Germany | 1,181 | 81.5 % |
| MY — Malaysia | 70 | 4.8 % |
| US — United States | 67 | 4.6 % |
| CH — Switzerland | 61 | 4.2 % |
| FR — France | 46 | 3.2 % |
| ES — Spain | 26 | 1.8 % |
| HU — Hungary | 20 | 1.4 % |
| IT — Italy | 10 | 0.7 % |
| PL — Poland | 9 | 0.6 % |
| GB — United Kingdom | 8 | 0.6 % |
| AT — Austria | 5 | 0.3 % |
| IE — Ireland | 4 | 0.3 % |
| SG, NL, CA, AU | 3 each | 0.2 % each |
| (no country code on the inventor record) | 708 | 48.9 % |
Data table: German regions (NUTS-1, EP applications only)
| NUTS-1 | Federal state | Families |
|---|---|---|
| DE1 | Baden-Württemberg | 426 |
| DE7 | Hessen | 325 |
| DE9 | Niedersachsen | 78 |
| DE2 | Bayern | 34 |
| DED | Sachsen | 29 |
| DEA | Nordrhein-Westfalen | 28 |
| DE3 | Berlin | 17 |
| DEG | Thüringen | 12 |
| DE4 | Brandenburg | 6 |
| DEB | Rheinland-Pfalz | 4 |
| DEF | Schleswig-Holstein | 4 |
| DE6 | Hamburg | 3 |
States with fewer than 3 families are omitted. 5,652 distinct inventor person records appear across the portfolio.
Methodology
Search strategy, data basis, and known limitations
Data basis of this edition
Recomputed on 15 August 2026 against
EPO PATSTAT Global, Spring 2026 (patstat-mtc.patstat_2026a).
Classification labels are taken from the classification views at
IPC edition 20260101 and CPC edition 202608. Every figure
in this report was produced by a query against that edition; the SQL is shipped with the
report in queries.sql.
Search strategy, family counting and data source — in detail
Defining the group
B. Braun is a family-owned group that files under many names, so the first decision is a boundary decision, not a data question. This report analyses the group: the parent B. Braun Melsungen AG together with the operating subsidiaries that file under a B. Braun name and with Aesculap AG, the wholly owned surgical-technology company in Tuttlingen. Aesculap shares no name element with the parent and would be invisible to any search anchored on the parent name, yet it is the largest single filer in the group. Excluding it would have described the parent company rather than the group the report is about.
Building the scope
-
The scope is built on
person_nameintls206_person, restricted to the applicant role (applt_seq_nr > 0). The harmonisedhan_namewas used only as a cross-check. - Discovery was deliberately broader than a prefix search: every applicant record whose name contains Braun or Aesculap anywhere, plus explicit searches for Chinese, Japanese and Korean renderings. That returned 145 applicant name records with at least one family in the period, each reviewed individually.
- Office-side renderings are included. Chinese 贝朗 (B. Braun) and 蛇牌 (Aesculap) and Japanese katakana forms such as ベー・ブラウン・メルズンゲン and ビー．ブラウン アビタム carry themselves as their own harmonised name and are reachable by no Latin-script search.
- Deliberately excluded as unrelated companies sharing the name stem: Braun GmbH / Braun AG and De’Longhi Braun Household (personal-care appliances, Kronberg), Hartmann & Braun, The Braun Corporation, Braun Elevator, TricorBraun, Asea Braun Boveri, Brauns-Heitmann, the Braunkohle and Braunschweig name families, individual inventors surnamed Braun, and Aesculapius Farmaceutici / Aesculapius Limited.
-
han_name cross-check, both directions. A scope built on
han_namereturns 1,447 families against the 1,450 of theperson_namescope: it adds 0 families (no contamination by a foreign organisation) and misses 3 (the non-Latin renderings). The real hazard here is not spelling but grouping — Aesculap records harmonise toAESCULAP AG, so a harmonised-name search anchored on the parent name alone reaches none of the 606 Aesculap families. -
Patents of invention only. All queries set
ipr_type = 'PI'. Without that filter, 191 further families first filed in the period would enter the portfolio — 136 of them utility-model filings only and 55 design filings only. Utility models are not marginal for a German applicant: a further 110 families inside the scope carry B. Braun utility-model applications alongside their patent applications; those families are in the report, their utility-model members are not counted as separate families.
Family counting and the year of a family
-
All counts are DOCDB patent families (
COUNT(DISTINCT docdb_family_id)), so an invention filed in ten countries counts once. - A family is dated by the earliest filing year over all its members — including members filed before 2014 and members filed by other applicants — and only then filtered to 2014–2024. A family first filed in 2012 and extended in 2015 is therefore outside this report rather than being counted in 2015. Under this definition the yearly figures add up exactly to the portfolio total of 1,450.
- The PCT rate counts families containing any WO member (737 of 1,450, 50.8 %), not the number of WO applications.
- Grant rates are computed per office on applications of the 2014–2020 filing cohort, so that pending examination does not depress the figure. WO is excluded because the PCT does not grant.
Multi-assignment
- A family normally carries several IPC symbols and several WIPO technology fields, assigned by different offices to different family members. This analysis aggregates all symbols from all members, so class and field totals sum well above 1,450 and are not shares of a whole. For the same reason no pie or doughnut chart is used anywhere in this report.
- Entity totals likewise exceed the portfolio total: the eight entity rows sum to 1,466 because 16 families are co-filed by two group entities.
- Office totals exceed the portfolio total because one family is normally filed at several offices.
Classification scope
- The three classification-scope checks — subtree expansion, predecessor union, retention of retired symbols — are not applicable to this report. The patent set comes from applicant harmonisation, not from a classification code list, so there is no code set to expand or complete. Classifications are used only to describe the set that the applicant search returned.
-
Class labels are joined from the classification views on
symbol_patstatand are read at the current editions (IPC 20260101, CPC 202608). Retired symbols carry no title in those views and are joined with an outer join so that they are labelled rather than dropped; no such symbol reached the reporting thresholds.
Data source and toolchain
- EPO PATSTAT Global, Spring 2026 (
patstat-mtc.patstat_2026a) on Google BigQuery. - Classification views
patstat-mtc.classification.dpt_ipcanddpt_edition: IPC 20260101, CPC 202608. - Tables used:
tls201_appln,tls206_person,tls207_pers_appln,tls209_appln_ipc,tls230_appln_techn_field,tls901_techn_field_ipc. - Access through the mtc.berlin PATSTAT MCP server; analysis and drafting with Claude (Opus 5).
Scope Limitations
- Group boundary by filing name. The group is defined by the applicant name recorded on the application. Companies acquired by B. Braun that continued to file under their previous name are therefore not included: a targeted check found 14 further families (0.97 % of the portfolio) filed 2014–2024 under Christoph Miethke GmbH & Co. KG (10) and Transcoject GmbH (4). PATSTAT records no ownership link and no acquisition date, so attributing them would be an assumption rather than a measurement. Filings under the name B. Braun Miethke GmbH are included (4 families).
-
Utility models and designs are excluded (
ipr_type = 'PI'): 136 utility-model-only and 55 design-only families first filed in the period are outside the report. - 2023 and 2024 are incomplete. Patent applications are normally published 18 months after priority, so families first filed in these years are still entering the database. Both years are shown but excluded from every trend statement.
- National-phase timing. PCT applications enter the national phase up to 30 months after priority. National office counts for families first filed in 2021–2022 are therefore lower bounds, which affects the direction — though not the existence — of the shift away from direct national filings.
- Inventor location is partial. 708 families contain at least one inventor record without a country code; NUTS regional data exists only for EP applications (979 of 1,070, 91.5 %) and not at all for German national filings.
- Grant data coverage. Offices that register rather than examine show rates at or near 100 % and are not comparable with examining offices; Singapore shows no grant information at all in this source.
- No external benchmark. All comparisons in this report are internal to the portfolio. No claim is made about B. Braun’s position relative to other medical technology companies, because no competitor set was measured.
Glossary — patent terms explained
- DOCDB patent family
- A group of patent applications protecting the same invention in different countries. Counting families instead of applications prevents an internationally filed invention from being counted many times.
- First filing year
- The earliest filing year among all members of a family — the year the invention entered the patent system.
- IPC / CPC
- International and Cooperative Patent Classification: hierarchical schemes that describe what a patent is about. A subclass such as A61B is coarse, a main group such as A61B 34/30 is specific.
- WIPO technology fields
- A mapping of IPC symbols onto 35 broad technology fields, used to compare portfolios across schemes and time.
- PCT / WO application
- An international application filed under the Patent Cooperation Treaty. It reserves the right to file in most countries of the world for up to 30 months, but is never granted itself.
- Applicant vs. inventor
- The applicant owns the application (normally the company); the inventor is the natural person. PATSTAT records both roles separately.
- han_name
- The harmonised applicant name supplied with PATSTAT, which groups spelling variants of an organisation. It is a helpful cross-check but is not an ownership model — subsidiaries with their own brand form their own group.
- Grant rate
- The share of applications at an office recorded as granted. It depends strongly on the office and on how much time has passed since filing.
Report files
This report is published as a set of source files: meta.json with the report
metadata, cover.html and kpi-strip.html, thirteen sections
under sections/, and queries.sql containing every SQL statement
used, with the PATSTAT and classification editions recorded in its header. Any figure in the
report can be traced to one of those statements and re-executed against the same edition.
Like what you see?
This report was built with a fully reproducible pipeline: EPO PATSTAT Global on BigQuery, a custom MCP server, and Claude AI for analysis and visualisation. Every number traces back to a query shipped with the report.