EU Pharmaceutical Reform: How the 2028-29 Filing Cutoff Changes Exclusivity

The EU pharmaceutical package is close to final adoption. The Council adopted its first-reading positions, ST-7105/26 and ST-7106/26, on 28 September 2026 [1], with 26 Member States in favour, none against and one abstention [2]. On 10 November, Parliament's SANT committee votes on its second-reading recommendations for both files, the directive (2023/0132) and the regulation (2023/0131) [3], and its chair has already told the Council that the committee will recommend approval without reopening the texts [2]. The plenary vote has been postponed twice, from early October to 19 October and then to the week of 23 November [4]. Once Parliament approves, the texts need only publication in the Official Journal to enter into force [1].

The package is usually described as a cut in EU exclusivity. Our reading of the adopted texts, the trial registries, the patent registers and the relevant financial filings leads to a narrower conclusion. The package changes the economics of when to file, whether to pursue extensions and how to meet supply obligations. Effective exclusivity changes less than is often assumed for products protected by patents, and more for orphan and paediatric programs. For every product, the filing date determines which rules apply. That follows from one sentence in the transitional provisions, Article 186(5) of the new Regulation:

The date of application is 24 months after entry into force, and procedures still pending at that date are completed under the old law [5]. An initial marketing authorisation application submitted before the cutoff therefore keeps the old protection periods, and one submitted afterwards falls under the new ones. Article 186(10) does the same for orphan medicines: products whose marketing authorisation application is submitted before the cutoff keep market exclusivity under Regulation 141/2000, for the initial marketing authorisation only [5].

Because the rule turns on a date, the programs it affects can be identified. We count them in the two trial registries, check them against the patent and SPC registers, and then estimate what the new incentives are worth, using a threshold model for the additional protection years a company can earn and a valuation of the transferable exclusivity voucher.

1. Filing before or after the cutoff

The adopted texts set out the transition explicitly. Pending centralised procedures are completed under Article 10 of Regulation 726/2004; Article 186(5) of the new Regulation keeps the old protection periods for reference products filed before the cutoff; and Article 186(10) keeps the old orphan exclusivity, for the initial authorisation only [5]. The table compares a filing in 2027 with a filing in 2029 for each type of asset [5][6][7][8][9][10].

Three rows need comment. Standard products lose one year of baseline protection, from 10 to 9, but can earn up to two years back. One additional year is available for addressing an unmet medical need or, for a new active substance, for trials against a comparator agreed in EMA scientific advice combined with filing first in the EU or with efficacy trials in more than one Member State. A second year requires a new indication with significant clinical benefit. A product that meets the conditions ends with the same protection as before or more [10]. The paediatric orphan row shows the largest change in the table, from 12 years to 9, because the new Regulation removes the reward for completing a paediatric investigation plan (PIP) [5][9]. The transitional provision for orphans keeps only the 10-year period of Regulation 141/2000 for applications filed before the cutoff, so it is not certain that those applications keep the two paediatric years either [5]. The voucher row is the only one that can exceed 11 years, and the number of vouchers is capped, as discussed in section 5 [5].

The economics literature adds a caveat: companies choose when to file. Danzon's study of 25 markets found that lower expected prices or smaller expected markets predicted fewer and later launches, so manufacturers already time their submissions to expected revenue [11]. The cutoff turns that gradual choice into a sharp break. A program that would normally file in the EU in 2029 now has a reason to consider filing before the cutoff, and section 4 estimates what that option is worth.

The old rules: 8+2 years, up to 11

Some transition presentations describe the old regime as "8+1". The legal texts say otherwise. Article 1(8) of Directive 2004/27/EC provides that a generic "shall not be placed on the market until ten years have elapsed" from the initial authorisation of the reference product, extended to a maximum of eleven years if a new indication with significant clinical benefit is authorised during the first eight years [6]. Article 14(11) of Regulation 726/2004 gives centrally authorised products "an eight-year period of data protection and a ten-year period of marketing protection", which can be extended to 11 years on the same conditions [7]. The correct comparison is therefore 10 years, up to 11, under the old rules, against 9 years, up to 11 or 12 with a voucher, under the new ones [5][6][7]. Paediatric orphan products received 12 years under the old rules [9]. A model that assumes the old regime was 8+1 understates what an earlier filing preserves, by one to three years per asset.

2. How many trials are affected

To estimate how many programs face the cutoff, we counted trials in the registries directly. The query covered every ClinicalTrials.gov trial classified as phase 2 or phase 3, including the registry's phase 1/2 and phase 2/3 categories, with an estimated primary completion date (PCD) on or after 1 January 2026 and at least one site in an EU-27 or EEA country. We retrieved all matching records from the registry API on 4 November 2026 and checked the total count [12]. The query returned 5,272 trials [12][13].

ClinicalTrials.gov does not hold every trial relevant to the EU. A search of the EU's Clinical Trials Information System (CTIS) for interventional phase 2/3 and phase 1/2 trials with active status returned 6,244 trials [14]. We matched these against all 67,469 phase 2/3 trials registered worldwide on ClinicalTrials.gov and updated since 2021, using shared asset names or codes, title similarity within the same sponsor, and title similarity across all records. The matching identified 3,473 of the 6,244 CTIS trials as also registered on ClinicalTrials.gov [13]. The remaining 2,771 trials, in 2,713 programs, are registered only in the EU and would be missed by a count based on ClinicalTrials.gov alone [15].

Combined, the trials that straddle the cutoff, which we call the straddle cohort, number 8,043 in 7,460 programs: 5,272 trials and 4,747 programs from ClinicalTrials.gov, and 2,771 trials and 2,713 programs found only in CTIS [13].

The CTIS-only trials come with two limitations. The CTIS search does not provide primary completion dates, so these 2,771 trials cannot be assigned to a completion year. The cross-registry matching is approximate, and we spot-checked it: in 12 audited records we found one trial classed as EU-only, EU CT 2026-526093-16-00 (pridopidine in Huntington's disease), that has a ClinicalTrials.gov record the matching had missed. On that basis, the number of CTIS-only trials lies between roughly 2,300 and 2,771 [15]. The totals above depend on both limitations.

Data: complete ClinicalTrials.gov API v2 retrieval, 4 November 2026, grouped by the sponsor's estimated primary completion year; CTIS-only trials from the CTIS search after cross-registry matching; see Sources.

At program level, the five completion-year groups contain 1,308, 1,278, 920, 600 and 641 programs, taking a program's completion date as the latest primary completion date among its trials [12]. Most of the cohort is active: 2,865 of the 5,272 ClinicalTrials.gov trials were recruiting on the retrieval date, and 144 had been first posted in the previous 90 days [12]. Many programs are global. Of the 5,272 trials, 3,043 (57.7 percent) also have sites outside the EU/EEA, so US FDA milestones and EU filing decisions often concern the same programs [12]. Classified by asset type from the condition and intervention fields, the ClinicalTrials.gov programs comprise 2,843 small molecules, 1,184 monoclonal antibodies, 321 cell or gene therapies (ATMPs), 246 ADCs, 94 GLP-1 or obesity programs, 68 vaccines, 52 radioligands and 420 other or unclassified programs; some programs fall into more than one class [12].

The orphan share

We estimated the orphan share by matching each program's active substance and indication against EMA's register of orphan designations (an export of 3,273 designation records generated on 4 November 2026). On a strict match of both substance and indication, 1,354 of the 4,747 ClinicalTrials.gov programs (28.5 percent) have a positive orphan designation; on a broader match by indication, 1,522 do (32.1 percent) [16]. Among the CTIS-only programs, 660 of 2,713 (24.3 percent) match strictly and 742 (27.3 percent) match broadly [16]. That puts roughly 2,000 to 2,300 orphan-linked programs in the cohort. The figure matters because the new rules change protection most for orphan products.

How completion delays affect the 2028 group

The 1,025 ClinicalTrials.gov trials due to complete in 2028 form the transition group: their results could support filings either before or after the cutoff. Estimated completion dates often slip, and the registry cannot show by how much, because a record's estimated date is replaced with the actual date once a trial completes, and earlier versions of a record are not available through the public API. We therefore modelled the delays as scenarios, and the scenario values are assumptions rather than findings. To measure how long delays last, we took all EU/EEA phase 2/3 trials on ClinicalTrials.gov with a primary completion date between January 2021 and October 2026, retrieved live on 4 November 2026, and found 1,809 whose estimated completion date had passed without an actual date being reported. The median overdue period was 20.6 months, with quartiles of 7.7 and 38.0 months [17]. If the share of trials that miss their estimated date (phi) is 0.25, 0.50 or 0.75, and the delays follow this distribution, then 822, 619 or 417 of the 1,025 trials still complete in 2028 and the rest move to 2029 or later [17]. Per 1,000 trials, the figures are 802, 604 and 406 [17]. Three caveats apply. Phi is an assumption, because the miss rate cannot be estimated from a single version of each record. The model assumes that the delay patterns of trials due between 2021 and 2026 will hold for trials due in 2028 [17]. And the shortest delays include records of trials that finished on time but have not yet been updated, which raises the number of trials that stay in 2028. On this distribution, a trial that misses its date has a 79 percent chance of moving out of 2028 [17].

Sponsors with programs completing in 2029 or later

We ranked lead sponsors, grouped by parent organisation, by the number of programs completing in 2029 or later, which are the programs that will file under the new rules. The top 25 groups hold 527 of the 1,241 such programs, or 42.5 percent [12][18]. Of those 527 programs, 169 belong to groups headquartered in the EU/EEA and 358 to groups headquartered elsewhere, a split of 32.1 to 67.9 percent; the groups outside the EU/EEA are mainly from the US, the UK, Switzerland and Japan [18]. The ranking is Merck & Co/MSD 71, AstraZeneca 51, Assistance Publique-Hopitaux de Paris 47, Novartis 36, Bristol Myers Squibb 30, AbbVie 26, Roche 21, Pfizer 21, Eli Lilly 20, UNICANCER 19, Gustave Roussy 19, Johnson & Johnson 19, Sanofi 15, Oslo University Hospital 14, Boehringer Ingelheim 13, Regeneron 13, University Hospital Rouen 12, Netherlands Cancer Institute 11, Gilead 11, Takeda 11, Daiichi Sankyo 10, Nantes University Hospital 10, GSK 9, Amgen 9 and University Hospital Toulouse 9 [12]. The registries do not record sponsor headquarters, so we classified them using a curated list of major companies and Wikidata; 388 ClinicalTrials.gov programs (8.2 percent) remain unclassified [18].

3. Patent and SPC protection compared with the 11-year cap

If the new 11-year cap set the end of protection for most products, the shorter baseline described in section 1 would mean an earlier loss of exclusivity for most of the programs counted in section 2. The patent registers show that it does not. We retrieved the EPO INPADOC and national SPC register records for the basic patents of a sample of 33 marketed products chosen to cover the main asset types. For 26 of them we verified the full chain from patent to SPC to market protection; for the other 7 we found no SPC [19]. In 24 of the 26 verified products, patent and SPC protection runs beyond the new 11-year cap, by +2.1 to +5.5 years in our summary of the register records [19]. This is consistent with the literature, which treats patents, SPCs and data exclusivity as separate layers of protection [20][21][22]; case studies of SPCs conclude that the certificates alone substantially delayed competition in Europe [22].

The verified rows are listed below, with the expiry dates and gaps as recorded in our register analysis.

No patent with an SPC found means our register search identified no SPC-protected core patent; where none exists, the regulatory protection period sets the end of exclusivity. Data: EPO INPADOC SPC register records for each basic patent and EMA EPAR authorisation dates, retrieved 3 to 4 November 2026. Gap figures are as recorded in our register analysis and were not recalculated. Dates vary by country because the SPC calculation can use the first national authorisation in the EEA, and paediatric extensions have been granted in some Member States only.

The summary range of +2.1 to +5.5 years does not match the individual rows, which run from +1.9 years (Soliris) to +5.3 years (Darzalex). We report both as recorded. Either range leaves 24 of the 26 verified products protected beyond the cap [19].

Kadcyla is the only verified product where the cap makes a difference: its composition patent expired in May 2024, about half a year before the 11-year cap would have ended protection in November 2024 [19][35].

Two rows need a closer look, because they show where the reform does change the outcome. Kymriah can appear to rely only on orphan exclusivity, because its basic patent was relinquished. The register shows otherwise. The relinquished patent, EP3214091B1, was a divisional. A related patent in the same family, EP2649086B1, granted to the Trustees of the University of Pennsylvania in 2017, has 26 granted SPCs in 20 countries running to August 2033, the limit of 15 years after authorisation. No opposition has been filed, and a UPC opt-out was registered in 2023. Patent and SPC protection therefore runs 4.1 years beyond the new-regime cap of August 2029. The earlier voluntary abandonment shows that this protection could change under opposition, but it was in force when we searched the register [47]. Zolgensma can appear to have no core patent with an SPC. In fact EP3505635B1 belongs to third parties, Genethon and CNRS, and Novartis's gene therapy unit is the licensee. Its 15 SPCs run to October 2033 in the countries where they remain in force, the French one having already been withdrawn, so the end of protection depends on licensing and settlement terms. It falls 2.4 years after the May 2031 cap [48].

Other cases point the other way. For first-generation monoclonal antibodies, biosimilar entry depended on development time rather than on patents: the first trastuzumab biosimilar was authorised 3.3 years after patent and SPC expiry, and the first rituximab biosimilar about 3.5 years after its patent expired (it had no SPC). Neither the old nor the new rules would have changed those dates [29][30]. New formulations can extend a product's commercial life regardless: in the Netherlands, subcutaneous trastuzumab reached a 50 percent share and settled near 20 percent after biosimilars entered, so the expiry date does not on its own determine market share [56]. Patents can also be lost early in litigation. Soliris's later composition patent was revoked or invalidated in the UK and at the UPC in 2024-25 [31], Xarelto's once-daily patent was revoked in the UK and Germany before expiry [24], and the SPC records for the salt form of Jakavi show the same pattern [27]. We did not search litigation dockets systematically, so litigation is included only where the registers showed that it changes which form of protection applies.

The seven rows without an SPC need a caveat as well. They rest on negative searches: we found no core patent with an SPC in the registers we searched, which does not prove that none exists in every national register. Where there is no patent protection, the regulatory periods decide. The two mRNA vaccines illustrate this. For Comirnaty and Spikevax, protection under the old rules ends in 2030 and 2031; under the new rules the baseline would end in 2029 and 2030, with the cap in 2031 and 2032. For such products, the period between the baseline and the cap is what sets the end of exclusivity [19].

These shares can be applied to the registry cohort only with care. In the register sample, regulatory protection applies to between 2 of 33 rows (the two mRNA vaccines, where it is certain) and 8 of 33 rows (the seven rows without an SPC plus Kadcyla), and patent protection applies to 24 of the 26 verified rows [19]. Applied as a range to the orphan-linked programs, the new rules could change the commercial exclusivity dates of roughly 122 to 488 programs on the strict orphan match, and up to 549 on the broad match, out of roughly 2,000 to 2,300 orphan-linked programs [19]. This is an estimate by asset type, not a product-by-product match. The verified sample is weighted towards marketed products with strong patent protection, and the patent families behind pipeline products are younger, so the share of pipeline trials protected mainly by patents is probably lower, and the share governed by regulatory protection probably closer to the top of the range [19].

4. What an additional year of protection is worth

Under the new rules, the 9-year baseline can rise to 11 years. One additional year can be earned in several ways: by addressing an unmet medical need or, for a new active substance, with trials against a comparator agreed in EMA scientific advice combined with filing first in the EU or with efficacy trials in more than one Member State. The other requires a new indication with significant clinical benefit [10]. For a product that does not qualify through unmet need, the comparator route depends on a spending decision. The company commits to an appropriately designed comparative study and, if the product is authorised and the extension is granted and kept, gains one more year of protection in the EU. The break-even point can be calculated from four inputs [57]:

C is the additional cost of the qualifying comparative study, and R* is the break-even value of one additional protected year, in euros of the extension year. p is the joint probability that the program reaches the extension year and that the extension is granted and kept, and Δt is the time between the midpoint of the study spending and the midpoint of the extension year. m is the contribution margin, and δ is the share of the extension year's contribution that generic or biosimilar entry would otherwise remove. All figures in this section come from our threshold model, and the table can be recalculated from the parameters given for each row [57].

Break-even value of the protected year's contribution, R*, in EUR M, at r = 10 percent and Δt = 12 years. Data: INBISTRA threshold model.

With an assumed margin of 70 percent and an assumed incremental capture of 65 percent, the C = 40 row corresponds to roughly 345 to 460 million euros of EU net sales at steady state [57].

In a worked example, all commercial figures are assumptions. A hypothetical specialty biologic that has completed phase 2 spends 45 million euros on the comparative study around year 1, files in year 3 with the comparative results in the dossier, launches in year 4, and reaches the extension year 12.5 years after the midpoint of the study spending. With assumed EU steady-state sales of 600 million euros, the extension year is worth 56.5 million euros in present value against a present-value cost of 40.9 million, a net gain of 15.6 million and a cover of 1.38 times [57]. The study stops paying for itself if any one of these holds: incremental capture of 0.47 or less, a joint probability of 0.543 or less, EU steady-state sales below 434 million, or more than 15.9 years between the spending and the payoff [57].

Two factors help explain why companies rarely make this investment. The first is organisational. The cost is certain, comes early and sits in the R&D budget, while the payoff comes a decade later in a commercial budget that may not yet exist, and teams tend to underinvest in that situation. Each additional year between the study and the extension raises the break-even value by 10 percent, from 122.5 million at an 8-year gap to 217.0 million at 14 years [57]. The second is structural. The old law already had a similar incentive: the additional year under Article 14(11) of Regulation 726/2004 for a new indication with significant clinical benefit. No public register records how often it has been used, and the register data suggest why. In the section 3 sample, 24 of 26 verified products are protected by patents and SPCs beyond the regulatory cap, so an eleventh year of regulatory protection had little commercial value, and the extension was of little use wherever patents set the end of exclusivity [57]. Under the new rules this changes. With a 9-year baseline and additional years to earn, the extra year is more likely to set the end of protection for products whose patents expire early, the kind of product represented by the rows without an SPC in section 3.

Two caveats limit these figures. The study cost range of 20 to 80 million euros is an assumption, based indirectly on median costs of US pivotal trials, because no EU data on the cost of comparative trials are available. EU authorities must also accept that the study was appropriately designed, so a disagreement over design can cost the extension after the money has been spent. For programs filing after the cutoff, the decision on the comparative study has to be made now, between 2026 and 2028, years before any revenue [57].

5. Valuing the transferable exclusivity voucher

The new Regulation creates a tradeable asset with no price history in Europe. The transferable exclusivity voucher, granted for an approved priority antimicrobial, extends regulatory data protection by 12 months for one centrally authorised product. It can be used for the antimicrobial itself or transferred once to another product and holder [5]. The scheme starts at entry into force, and the first vouchers can only be granted once a priority antimicrobial has been approved, so any valuation has to be built from first principles [5]. Using the income method, with the margin and erosion assumptions from section 4, we value one voucher at EUR 52 to 253 million of income value under the sales cap, and at EUR 40 to 220 million in an actual transfer, with a central range of EUR 100 to 165 million [58]. The upper end is set by a legal limit. A buyer can use the voucher only in the fifth or sixth year of the receiving product's data protection, and only if the product's audited annual gross EU sales were at or below EUR 490 million in any of its first four years [5]. A company that uses the voucher for its own antimicrobial faces no sales cap, but EU antimicrobial sales are small, so self-use is worth roughly 3 to 12 times less than a transfer [58]. The scheme also makes prices public: the buyer must report the transfer value to the Agency within 30 days, and the Agency publishes it [5]. The first transfer will therefore establish a market price.

US sales of priority review vouchers, which we checked against SEC filings, provide reference points. BioMarin sold a rare pediatric disease priority review voucher to Regeneron Ireland for USD 67.5 million in July 2014, and a second voucher, for Brineura, in December 2017 for USD 125.0 million. PTC Therapeutics sold the Kebilidi voucher for USD 150.0 million under an agreement of November 2024, about EUR 82 to 142 million in 2026 terms [58][59]. Academic work attributes most of the value to scarcity. About a third of a voucher's commercial value comes from the market share it captures, half from earlier sales and less than a quarter from longer effective exclusivity, and the modelled value exceeds USD 200 million only if about one voucher is issued per year, falling below USD 100 million at four per year [60]. The EU scheme took transferability from the original proposal for such vouchers [61] and limited the number to five in fifteen years, which is at the scarce end of past schemes [5].

In a program valuation, the voucher can decide whether the program goes ahead. Our representative priority antimicrobial, with an illustrative upper-mid set of success probabilities, has a risk-adjusted NPV of minus EUR 11.99 million without the voucher. With the voucher included at a 0.6 probability of being granted, the NPV becomes plus EUR 0.67 million, an increase of 12.66 million in present value, and a voucher worth about EUR 114 million would on its own be enough for the program to break even [58]. The increase grows with the voucher's value and with the probability of obtaining it [58]. That probability depends on the conditions in Article 41(4), supply capacity, disclosure of funding and filing in the EU first, within 180 days [5].

Four risks affect the value. First, the US vouchers are an imperfect comparison: a priority review voucher speeds up FDA review [61], while the EU voucher adds a full year of protection [5]. The November 2024 sale for USD 150 million is still the closest reference for a first EU price. Second, the scheme is politically exposed. The adopted text limits it to five vouchers [62], and a first transfer price at the top of the US range would attract attention and could lead to a review of the scheme. Third, vouchers can be revoked. The Commission may revoke a voucher before transfer if supply or procurement requests are not met, and a voucher expires five years after it is granted, which could force a developer short of cash to sell within that period [5]. Fourth, the adopted texts do not settle how a voucher applies to a receiving product still covered by the old rules, so this needs confirmation. Buyers are also not limited to the top 20 companies: the voucher can be used for the priority antimicrobial itself or for another centrally authorised product of the same or a different marketing authorisation holder [5].

The evidence on whether vouchers encourage innovation is less favourable. Goh's review of 18 years of the FDA's neglected tropical disease voucher program found that the 14 vouchers awarded up to 2024 mostly went to filings for long-established therapies, with approval a median of 8.7 years after first approval or endemic use; the vouchers created transfer value without new R&D [63]. Berner-Rodoreda's assessment of the proposed EU vouchers reached a similar conclusion: they would keep prices high, delay competition and shift the cost of antibiotic development onto arbitrage rather than invention [64]. In the Commission's consultation of 478 respondents, industry favoured additional incentives such as transferable vouchers on top of existing IP protection, while other stakeholders preferred scientific advice and faster regulatory pathways [65]. A buyer pricing the first transfer should treat these findings as political risk, which the upper end of the range above already reflects.

6. Conditions attached to protection: supply, orphan exclusivity and the voucher cap

Three provisions make exclusivity depend on conditions. They are risks written into the law and say nothing about how any particular company will behave.

The supply obligation is new, and its reach can be measured. Under Article 59 of the new Directive, a Member State may ask a marketing authorisation holder to place a protected product on its market and supply enough to meet patients' needs, and may make that request within one year of authorisation. If the product has not been continuously available within three years of the request, the market protection and any orphan extension stop applying in that Member State, the sanction is published, and generic or biosimilar applications may be validated and assessed there from six years after the start of the reference product's data protection [5]. When the new rules take effect, at least 1,564 centrally authorised products will be subject to the supply obligations [66]. The notification deadlines are strict. Permanent cessation or withdrawal must be notified at least 12 months before the last supply, and a suspension or disruption of more than two weeks at least 6 months in advance. Prescription products need shortage prevention plans, including products authorised before the reform and brought in through Article 186(1) of the Regulation [5]. The Commission can fine breaches up to 5 percent of Union turnover and impose daily penalty payments of up to 2.5 percent of average daily Union turnover for the obligations in Annex II(26) [5]. For context, EMA received 104, 120 and 103 marketing authorisation applications in 2023, 2024 and 2025, and 45 in the first half of 2026; orphan applications numbered 23 in 2024, 21 in 2025 and 7 in the first half of 2026 [67]. The EMA shortages catalogue, which the new rules will manage, held 85 records on the retrieval date, 27 of them (32 percent) ongoing [68]. The first Union list of critical medicines, with more than 300 medicines and more than 200 active substances, indicates where the supply obligations are likely to apply first [69]. An exclusivity model now needs a supply plan for each Member State, alongside the pricing and ethical questions. The number of nationally authorised products across the EU is not published in aggregate, so national exposure can be bounded but not counted.

Holding several orphan authorisations for one substance no longer extends exclusivity. Where a holder has more than one orphan marketing authorisation for the same active substance, the authorisations no longer receive separate exclusivity periods. A single period runs from the first orphan authorisation in the Union, and each qualifying new orphan indication adds 12 months, at most twice, if granted at least two years before the end of the current period (Articles 73(3) and 74 of the new Regulation) [5]. The transition for this rule is narrow. A second orphan authorisation obtained before entry into force, expected around the end of 2026 or early 2027, can still receive a full 10-year period under the old rules, while one applied for but not granted before the cutoff falls under the single period [5]. Two further changes apply to orphan products. Applications for similar products may be submitted, validated and assessed during the last two years of orphan exclusivity, so competitors can enter as soon as it ends, and the six-year profitability review has been removed [5]. The changes by category (non-breakthrough orphans from 10 to 9 years, paediatric orphans relying on the PIP reward from 12 to 9, and 4 years for well-established-use orphans) apply in addition to the Article 186(10) transition for applications filed before the cutoff [5][8][9].

The voucher scheme is small. The adopted text limits it to five vouchers within 15 years of entry into force. Each adds 12 months of data protection, can be used for the priority antimicrobial itself or for another centrally authorised product, and can be transferred once, subject to the sales cap discussed in section 5 [5][62]. With only five vouchers available, an antimicrobial business case should treat obtaining one as uncertain, and should allow for revocation if supply and procurement requests are not met [5].

7. US pricing and tariff pressure

The package will take effect while US pricing and tariff policy is changing the conditions for launches. As of October 2026, heading 9903.04.62 under Section 232 applies an additional 15 percent duty to patented pharmaceuticals from the EU from 31 July 2026. It follows the April 2026 proclamation that set a default rate of 100 percent, with a zero rate until 20 January 2029 for companies that sign agreements on most-favoured-nation (MFN) pricing and onshoring [70]. MFN pricing is now the baseline: twenty-six large and mid-sized companies have signed voluntary MFN agreements reported to cover 89 percent of the branded market, and the proposed GLOBE and GUARD frameworks would work as mandatory international reference pricing [71]. At the informal trade council in Dublin on 22-23 October, the strategic dialogue on life sciences produced discussions but no joint text [72].

The available data point in both directions. The European pharmaceutical market was worth about EUR 330 billion at ex-factory prices in 2025, up from EUR 302 billion in 2024, and Europe accounted for 23.7 percent of world pharmaceutical sales in 2025, against 54.7 percent for North America [73]. Of 104 new molecules first launched worldwide in 2025, 16 came from European companies, compared with 46 from China and 28 from the US [73]. Of sales of medicines launched between 2020 and 2024, 74.1 percent were in the US and 15.6 percent in Europe's five largest markets [73]. Studies of launch timing show a structural gap: the first launch in the four largest EU markets or the UK comes on average about 10 months after US availability, most often in Germany, and the gap is growing [74]. Over the previous two decades European launch delays had fallen, from 37.2 months in 2000 to 11.8 months in 2017, so the recent widening reverses an earlier improvement [75]. Access has also worsened. The median time from central marketing authorisation to availability is 532 days, ranging from 56 days in Germany to 1,201 days in Romania, with an EU average of 597 days; the share of medicines fully available fell to 28 percent from 42 percent in 2019, and an independent analysis by the OHE confirms the decline [76][77]. Among new active substances approved by the FDA, the share also approved by EMA fell to 38 percent for the 2025 cohort from 62 percent for 2021, according to the EFPIA Regulatory Approval Indicator; this figure comes from a single table and should not be read as a headline trend [78].

The industry set out its own figures in an open letter on 22 September 2026. The chairs of AstraZeneca, Boehringer Ingelheim, Chiesi, Ipsen, GSK, Novo Nordisk, Novartis, Roche and Sanofi wrote that Europe's share of global pharmaceutical R&D had fallen from 43 percent in 1990 to 31 percent and was still falling, that its share of commercial clinical trials had halved in a decade to 9 percent, that China had overtaken Europe in trials, patents and new-medicine development, and that about 40 percent of newly approved therapies never reach European patients, while those that do take about 600 days [79]. EFPIA's response to the package was more critical. It called the package "not strong enough to move the needle on European competitiveness", citing the loss of a quarter of Europe's share of global investment over two decades and a halved share of trials [80].

These are risks, not findings about what any company will do. A launch model should include the tariff, the MFN baseline and the measurement caveats together, because the EU filing decision discussed in sections 1 and 2 is now made alongside a US market that usually launches first and sets prices under different rules.

8. Implications for launch models

For every asset headed for the EU, the reform adds several items to the launch model. The first is a scenario for filing before or after the cutoff, because filings in 2027 and 2029 fall under different rules and the difference is one to three years per asset. The second is patent and SPC expiry checked against the registers, because effective exclusivity ends at the later of that expiry and the regulatory cap, and the registers show which one applies. The third is a break-even threshold for the extension, set in phase 2 using the four inputs in section 4, so that the decision on a comparative study is costed before the filing deadline. The fourth is a voucher line with its own probability (section 5), with a check against double counting: the expected value of a voucher and the voluntary subscription model are separate incentives, and a program model should not count both as revenue [5]. The fifth is a supply plan for each Member State that treats Article 59 as a pricing risk, because three years without supply in one market now costs exclusivity there [5]. At portfolio level, the package changes valuations. Advisers are already telling holders to take it into account in current transactions and to reassess the value of EU manufacturing capacity under the Critical Medicines Act and the supply rules [81]. The reform's aim of encouraging launches across the EU also runs against the variety of national pricing and reimbursement systems, a trade-off the access literature has pointed out [82].

The table sets out the main protection periods by jurisdiction, each with its legal basis. It also corrects a common error: US exclusivity for the first interchangeable biosimilar is 1 year, 18 months or 42 months depending on the trigger, and the "4-year interchangeability" shorthand confuses it with the 4-year filing bar [83].

9. Conclusions

1. The filing date decides which rules apply. Applications submitted before the cutoff keep the old protection of 10 years, up to 11; later applications get 9 years, up to 11, or 12 with a voucher [5][10]. The cutoff is 24 months after entry into force, which on the current timetable means the end of 2028 or early 2029 [4][5].
2. Many programs are close to the cutoff. 8,043 phase 2/3 trials in 7,460 programs with EU/EEA sites complete from 2026 onwards, 1,025 of them in 2028, and between 417 and 822 of those are expected to still complete in 2028 [13][17].
3. For products with strong patents, the cut from 10 to 9 years makes little difference: patent and SPC protection outlasts the new 11-year cap in 24 of 26 verified products. The new periods decide the end of exclusivity where patents are weak or absent, which could affect roughly 122 to 549 of the 2,000 to 2,300 orphan-linked programs [16][19].
4. Only one extra year can be earned through unmet need, comparator trials or multi-country trials, and a second only through a new indication with significant clinical benefit [10]. A EUR 40 million comparator study pays for itself only above roughly EUR 345 to 460 million of EU sales, and the decision has to be taken in phase 2, with EMA advice and before EMA has published its guidelines [10][57].
5. The transferable exclusivity voucher is scarce and valuable, at about EUR 40 to 220 million in a transfer, and it can decide whether an antimicrobial program goes ahead. The evidence that vouchers produce new R&D is weak [58][63].
6. Protection now depends on supply. A Member State can remove market protection if a product is not supplied there within three years of its request, from 12 months after entry into force, and shortage prevention plans apply to products already on the market [5][10]. Delaying EU launches to protect US prices under most-favoured-nation pricing now carries an EU cost [10][71].
7. EU exclusivity can no longer be modelled as a fixed term. Deal terms should assume conservative protection with payments if the extra years are earned, and generic competition will start faster once protection ends because of the wider Bolar exemption [89][90].

The open points will be settled by three events: publication in the Official Journal, EMA's guidelines on unmet medical need and comparators, and the first published voucher price [5][10].

10. Methods

This article combines seven analyses, carried out in parallel. A review of the legal texts checked the old baselines and the adopted transition, supply and voucher provisions against the EUR-Lex consolidated texts and the Council positions ST-7105/26 and ST-7106/26. A patent register analysis retrieved the EPO INPADOC SPC records for each basic patent and checked them against EMA authorisation dates. A clinical trials analysis ran the full ClinicalTrials.gov query and the CTIS cross-registry matching, and a second analysis of live ClinicalTrials.gov records calibrated completion-date delays. A valuation analysis built the threshold model and the voucher valuation and verified both in code. Two further reviews covered web sources for the financial figures and the academic literature. Data extraction and cross-checking used ARiDA, INBISTRA's multi-agent analysis platform, working from the ClinicalTrials.gov API v2, the CTIS search, the EPO INPADOC and national SPC registers, EUR-Lex, EMA registers and exports, and SEC EDGAR filings. Every figure in the article comes from one of these analyses.

Section 2 describes the cohort query, the phase definitions, the rule for grouping trials into programs and the grouping by completion year, as they were run [12][13][15]. CTIS-only trials are not assigned to a completion year because CTIS does not publish completion dates. We did not search EudraCT, the register used under the old Directive. Trials that are still active are registered in CTIS and are included there, while EU-only trials registered before 2022 that have completed enrolment fall outside the active cohort.

The delay model uses live ClinicalTrials.gov API v2 records retrieved on 4 November 2026: the 11,303 EU/EEA phase 2/3 trials with a primary completion date between January 2021 and December 2028, found with the same query as the cohort and an earlier start date. Each record shows only its current estimated or actual date, so the gap between the first estimate and the actual date cannot be measured, and trials that moved their estimate forward before it passed do not appear in the delay distribution at all. Records whose estimated date has passed are counted as overdue; some belong to trials that finished on time but have not yet been updated, which shortens them. The miss rates are assumptions, and the results depend on the assumption that delay patterns are stable over time [17].

In the register analysis, SPC duration is the first EEA authorisation date minus the patent filing date minus five years, capped at five years, with an overall limit of 15 years from authorisation, plus six months for a paediatric extension where granted; dates vary between countries by 6 to 12 months. A missing SPC is a flag rather than proof, because INPADOC coverage differs between national offices and several SPC filings from 2024-2026 appear as filed but not granted, with grant status varying by country. For Japan, the specific investigation periods are set in an MHLW Cabinet Order that we did not retrieve; the table relies on the text of the PMD Act. Twelve rows use records retrieved on 3 November and were not retrieved again. We did not search litigation dockets systematically. The link between the register sample and the cohort is by asset type; we did not match individual trials to products [19].

Priority review voucher prices are taken from SEC filings [59]. Revenue ranges for advanced therapies are approximate, because EU revenue disclosures for these products are sparse.

All quotations of the new rules are from the Council's adopted first-reading positions. The Official Journal texts had not been published as of 4 November 2026, and Parliament's second reading was expected in the week of 23 November, so article numbers and dates may change in the published texts. The indicative plenary date of 23 November comes from the OEIL procedure file as reported by the Regunaut tracker; the OEIL page we retrieved directly shows the second-reading status but not the date. The earlier milestones were the provisional agreement on 11 December 2025, COREPER endorsement on 6 March 2026, SANT approval on 18 March 2026 and Council adoption on 28 September 2026 [62]. Where the evidence is incomplete, the text says so, and the remaining figures are labelled as assumptions.

Sources

1. [1] Council of the EU: https://www.consilium.europa.eu/en/press/press-releases/2026/09/28/pharma-package-council-adopts-new-rules-for-a-fairer-and-more-competitive-eu-pharmaceutical-sector
2. [2] PwC Switzerland: https://www.pwc.ch/en/insights/tax/pharma-life-sciences/eu-pharmaceutical-reform-time-to-rethink-your-operating-model.html
3. [3] European Parliament (doceo): https://www.europarl.europa.eu/doceo/document/SANT-OJ-2026-11-09-1_EN.html
4. [4] regunaut.com: https://www.regunaut.com/eu-directive/timeline
5. [5] data.consilium.europa.eu: https://data.consilium.europa.eu/doc/document/ST-7105-2026-INIT/en/pdf
6. [6] eur-lex.europa.eu: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32004L0027
7. [7] eur-lex.europa.eu: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:02004R0726-20070126
8. [8] eur-lex.europa.eu: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32000R0141
9. [9] eur-lex.europa.eu (Regulation (EC) No 1901/2006 on medicinal products for paediatric use, Article 37): https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32006R1901
10. [10] data.consilium.europa.eu: https://data.consilium.europa.eu/doc/document/ST-7106-2026-INIT/en/pdf
11. [11] The impact of price regulation on the launch delay of new drugs--evidence from twenty-five major markets in the 1990s.: https://doi.org/10.1002/hec.931
12. [12] ClinicalTrials.gov API v2, advanced query: phase 2 or phase 3, primary completion date from 1 January 2026, at least one location in an EU/EEA country (30 country terms), retrieved 4 November 2026: https://clinicaltrials.gov/data-api/api
13. [13] INBISTRA analysis: merged ClinicalTrials.gov and CTIS straddle-cohort program table with cross-registry deduplication, 4 November 2026
14. [14] EU Clinical Trials Information System (CTIS) public search, interventional phase 2/3 and 1/2 trials with active status (totals of 5,419 and 825 across the two filtered searches), retrieved 4 November 2026: https://euclinicaltrials.eu/ctis-public
15. [15] CTIS public trial records, cross-check of 12 sampled trials, including EU CT 2026-526093-16-00: https://euclinicaltrials.eu/ctis-public/view/2026-526093-16-00
16. [16] EMA orphan designations report (XLSX, generated 4 November 2026; 3,273 rows, 2,200 positive designations), matched by INBISTRA to program substance and indication text; CTIS orphan flags from trial records: https://www.ema.europa.eu/en/documents/report/medicines-output-orphan_designations-report_en.xlsx
17. [17] INBISTRA analysis: completion-date delay calibration on live ClinicalTrials.gov API v2 records of EU/EEA phase 2/3 trials with primary completion dates from January 2021, retrieved 4 November 2026: https://clinicaltrials.gov/data-api/api
18. [18] INBISTRA analysis: lead-sponsor grouping and headquarters classification (curated company map, Wikidata queried 4 November 2026, CTIS and ClinicalTrials.gov record fields): https://www.wikidata.org
19. [19] INBISTRA analysis: 33-product register sample (EPO INPADOC and national SPC registers, EMA EPAR authorisation dates), retrieved 3 to 4 November 2026, and its archetype-level transfer to the straddle cohort
20. [20] Evergreening, patent challenges, and effective market life in pharmaceuticals.: https://doi.org/10.1016/j.jhealeco.2012.01.004
21. [21] Data exclusivity exceptions and compulsory licensing to promote generic medicines in the European Union: A proposal for greater coherence in European pharmaceutical legislation.: https://doi.org/10.1186/s40545-017-0107-9
22. [22] Supplementary protection certificates and their impact on access to medicines in Europe: case studies of sofosbuvir, trastuzumab and imatinib.: https://doi.org/10.1186/s40545-019-0198-6
23. [23] Bristol-Myers Squibb Company (with Pfizer alliance): Patent EP1427415B1: https://register.epo.org/application?number=EP02775843&lng=en&tab=federated
24. [24] Bayer Pharma AG: Patent EP1261606B1: https://register.epo.org/application?number=EP00993610&lng=en&tab=federated
25. [25] Bristol-Myers Squibb Company: Patent EP1506211B1: https://register.epo.org/application?number=EP03736643&lng=en&tab=federated
26. [26] Gilead Pharmasset LLC: Patent EP2203462B1: https://register.epo.org/application?number=EP08732818&lng=en&tab=federated
27. [27] Incyte Corporation / Incyte Holdings Corporation (Novartis EU licensee): Patent EP1966202B1: https://register.epo.org/application?number=EP06839328&lng=en&tab=federated
28. [28] AbbVie Biotechnology Ltd (originally BASF/Knoll/Cambridge Antibody Technology): Patent EP0929578B1: https://register.epo.org/application?number=EP97906572&lng=en&tab=federated
29. [29] Genentech, Inc.: Patent EP0590058B1: https://register.epo.org/application?number=EP92914220&lng=en&tab=federated
30. [30] IDEC Pharmaceuticals/Genentech/Roche: https://pmc.ncbi.nlm.nih.gov/articles/PMC7188399
31. [31] Alexion Pharmaceuticals, Inc.: Patent EP0758904B1: https://register.epo.org/application?number=EP95919041&lng=en&tab=federated
32. [32] Merck Sharp & Dohme B.V. (Schering/Organon lineage): Patent EP2170959B1: https://register.epo.org/application?number=EP08768485&lng=en&tab=federated
33. [33] Regeneron/Sanofi: Patent EP2356151B1: https://register.epo.org/application?number=EP09744292&lng=en&tab=federated
34. [34] Genmab A/S / Janssen Biotech, Inc.: Patent EP2567976B1: https://register.epo.org/application?number=EP12178778&lng=en&tab=federated
35. [35] Genentech, Inc. (Roche): Patent EP2281577B1: https://register.epo.org/application?number=EP10011901&lng=en&tab=federated
36. [36] Daiichi Sankyo Company, Limited: Patent EP3101032B1: https://register.epo.org/application?number=EP15743738&lng=en&tab=federated
37. [37] Novo Nordisk A/S: Patent EP1863839B1: https://register.epo.org/application?number=EP06725149&lng=en&tab=federated
38. [38] Eli Lilly and Company: Patent EP3242887B1: https://register.epo.org/application?number=EP16703620&lng=en&tab=federated
39. [39] Eli Lilly and Company: Patent EP1641823B1: https://register.epo.org/application?number=EP04752589&lng=en&tab=federated
40. [40] Micromet Inc. (inventorship/ownership lineage; applicant Amgen Research (Munich) GmbH): Patent EP1691833B1: https://register.epo.org/application?number=EP04803302&lng=en&tab=federated
41. [41] Chugai Seiyaku Kabushiki Kaisha (Roche): Patent EP2644698B1: https://register.epo.org/application?number=EP11842145&lng=en&tab=federated
42. [42] Genmab A/S (Janssen Biotech ownership transfers recorded on the SPCs): Patent EP3337824B1: https://register.epo.org/application?number=EP16754148&lng=en&tab=federated
43. [43] Genmab A/S (Janssen Biotech development): Patent EP3487882B1: https://register.epo.org/application?number=EP17745951&lng=en&tab=federated
44. [44] Genmab A/S / Janssen Biotech, Inc.: Patent EP2922872B1: https://register.epo.org/application?number=EP13856592&lng=en&tab=federated
45. [45] Novartis AG (AAA lineage); ligand family from Ludwig-Maximilians-Universitaet Muenchen (Wester/Schottelius-lineage 2014 EP application; DE ref doc 602014089438): Patent EP4095130B1: https://register.epo.org/application?number=EP22178736&lng=en&tab=federated
46. [46] Legend Biotech Ireland Limited / Janssen Biotech, Inc.: Patent EP4063397B1: https://register.epo.org/application?number=EP22168161&lng=en&tab=federated
47. [47] The Trustees of the University of Pennsylvania (June, Levine, Porter et al.): Patent EP2649086B1: https://register.epo.org/application?number=EP11846757&lng=en&tab=federated
48. [48] Genethon (AFM-Telethon) / CNRS - THIRD-PARTY basic patent holders; Novartis Gene Therapies (AveXis) is the MA holder/licensee: Patent EP3505635B1: https://register.epo.org/application?number=EP19158152&lng=en&tab=federated
49. [49] BioNTech SE / Pfizer Inc.: https://www.ema.europa.eu/en/medicines/human/EPAR/comirnaty
50. [50] ModernaTX, Inc.: https://www.ema.europa.eu/en/medicines/human/EPAR/spikevax
51. [51] Advanced Accelerator Applications (AAA) / Novartis: Patent EP1598366B1: https://www.ema.europa.eu/en/medicines/human/EPAR/lutathera
52. [52] Immunocore Limited: https://www.ema.europa.eu/en/medicines/human/EPAR/kimmtrak
53. [53] Kite Pharma (Gilead): https://www.ema.europa.eu/en/medicines/human/EPAR/tecartus
54. [54] Juno Therapeutics (BMS): https://www.ema.europa.eu/en/medicines/human/EPAR/breyanzi
55. [55] Celgene (BMS) / bluebird bio (NCI-licensed): https://www.ema.europa.eu/en/medicines/human/EPAR/abecma
56. [56] The impact of an 'evergreening' strategy nearing patent expiration on the uptake of biosimilars and public healthcare costs: a case study on the introduction of a second administration form of trastuzumab in The Netherlands.: https://doi.org/10.1007/s10198-023-01648-w
57. [57] INBISTRA analysis: threshold model for the comparative-data extension year (r = 10 percent; row parameters as stated in section 4)
58. [58] INBISTRA analysis: transferable exclusivity voucher valuation (income method, transfer-value band, USD-to-EUR conversion in 2026 terms, representative priority-antimicrobial risk-adjusted NPV)
59. [59] SEC primary filings: BioMarin 8-K (USD 67.5M to Regeneron Ireland, closed 30 July 2014); BioMarin 10-K FY2018 (USD 125.0M, December 2017, Brineura PRV); PTC Therapeutics 8-K (USD 150.0M agreement 26 November 2024; net proceeds USD 148.0M per FY2025 10-K): https://www.sec.gov/Archives/edgar/data/1048477/000119312514291230/d769619d8k.htm
60. [60] The Commercial Market For Priority Review Vouchers.: https://doi.org/10.1377/hlthaff.2015.1314
61. [61] Developing drugs for developing countries.: https://doi.org/10.1377/hlthaff.25.2.313
62. [62] Legal 500 / CMS (Staber & Hofbauer): https://www.legal500.com/intelligence/european-union/food-drugs-healthcare-life-sciences/eu-pharmaceutical-package-what-the-councils-revised-texts-mean-for-companies
63. [63] Priority without progress: the FDA's neglected tropical disease voucher program after 18 years.: https://doi.org/10.1093/haschl/qxag024
64. [64] Transferable data exclusivity vouchers are not the solution to the antimicrobial drug development crisis: a commentary on the proposed EU pharma regulation.: https://doi.org/10.1136/bmjgh-2023-014605
65. [65] Unmet medical needs definition and incentives: stakeholders perspectives on the reform of the EU pharmaceutical legislation.: https://doi.org/10.3389/fmed.2024.1506243
66. [66] INBISTRA analysis of EMA medicines data (centrally authorised human medicines), retrieved 4 November 2026: https://www.ema.europa.eu/en/medicines/download-medicine-data
67. [67] European Medicines Agency, Monthly statistics report: June 2026, medicinal products for human use (EMA/160125/2026): https://www.ema.europa.eu/en/documents/report/medicinal-products-human-use-monthly-figures-june-2026_en.pdf
68. [68] European Medicines Agency, shortages catalogue (public information on medicine shortages), retrieved 4 November 2026: https://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/medicine-shortages-availability-issues/public-information-medicine-shortages
69. [69] European Medicines Agency, Union list of critical medicines: https://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/medicine-shortages-availability-issues/availability-medicines-during-crises/union-list-critical-medicines
70. [70] customsgenius.com: https://www.customsgenius.com/articles/9903-04-62-section-232-pharmaceuticals-15-percent
71. [71] pharmexec.com: https://www.pharmexec.com/view/mfn-offer-cant-refuse
72. [72] gov.ie: https://www.gov.ie/en/department-of-foreign-affairs/press-releases/minister-mcentee-to-host-meeting-of-trade-ministers-in-dublin
73. [73] EFPIA (The Pharmaceutical Industry in Figures — Key Data 2026): http://www.efpia.eu/media/owqczcqz/the-pharmaceutical-industry-in-figures-2026.pdf
74. [74] IQVIA (EMEA Thought Leadership): https://www.iqvia.com/locations/emea/blogs/2026/07/tracking-mfns-impact-on-innovative-medicine-launches
75. [75] Changes in launch delay and availability of pharmaceuticals in 30 European markets over the past two decades.: https://doi.org/10.1186/s12913-022-08866-7
76. [76] efpia.eu: https://www.efpia.eu/news-events/the-efpia-view/statements-press-releases/patients-in-europe-waiting-longer-for-new-medicines-as-inequality-grows-between-member-states
77. [77] Office of Health Economics (OHE): https://www.ohe.org/insights/how-long-are-patients-waiting-for-new-treatments
78. [78] EFPIA / IQVIA (Regulatory Approval Indicator, quarterly): https://www.efpia.eu/media/0l4nxwup/efpia-regulatory-approval-indicator-quarterly-qtr1-2026.pdf
79. [79] Novartis (hosting open letter by the Chairs of AstraZeneca, Boehringer Ingelheim, Chiesi, Ipsen, GSK, Novo, Novartis, Roche, Sanofi): https://www.novartis.com/news/europe-losing-pharma-investment-race-comeback-within-reach-open-letter-from-chairs-astrazeneca-boehringer-ingelheim-chiesi-group-ipsen-gsk-novo-novartis-roche-and-sanofi
80. [80] EFPIA (statement): https://www.efpia.eu/news-events/the-efpia-view/statements-press-releases/efpia-response-to-pharmaceutical-package-agreement-more-to-do-to-keep-european-pharma-competitive
81. [81] twobirds.com: https://www.twobirds.com/en/trending-topics/eu-pharmaceutical-package
82. [82] From Vision to Reality: The EU's Pharmaceutical Reforms and the Path to Improved Access.: https://doi.org/10.1007/s41669-024-00556-w
83. [83] govinfo.gov: https://www.govinfo.gov/link/uscode/42/262
84. [84] govinfo.gov: https://www.govinfo.gov/link/uscode/21/355
85. [85] govinfo.gov: https://www.govinfo.gov/link/uscode/42/1320f-1
86. [86] legislation.gov.uk: https://www.legislation.gov.uk/uksi/2012/1916/regulation/51B
87. [87] laws.e-gov.go.jp: https://laws.e-gov.go.jp/law/335AC0000000145
88. [88] laws.e-gov.go.jp: https://laws.e-gov.go.jp/law/334AC0000000121
89. [89] Morgan Lewis, The New EU Pharmaceutical Legislation: Key Points for Deal Teams, 22 June 2026: https://www.morganlewis.com/pubs/2026/06/the-new-eu-pharmaceutical-legislation-key-points-for-deal-teams
90. [90] Osborne Clarke, EU pharma package widens the Bolar exemption and moves the line on patent infringement, 2 November 2026: https://www.osborneclarke.com/insights/eu-pharma-package-widens-bolar-exemption-and-moves-line-patent-infringement