Breaking a drug patent: what Brazilian law authorises and what it does not solve

By Ramon Martins Andrade, attorney (OAB/RJ 188.374)

I am writing on the closing day of AIDS 2026, the 26th International AIDS Conference, held at Riocentro in Rio de Janeiro from 26 to 31 July under the theme Rethink. Rebuild. Rise. The opening ceremony was interrupted by protesters whose signs accused the pharmaceutical industry of greed and denounced the use of their bodies in research followed by the denial of access to its results. Brazilian activists brought a manifesto demanding that lenacapavir be purchased at a price the public health system can sustain and that the country confront the patent monopoly. One phrase circulated in the corridors all week, in Portuguese and in English: break the patent.

The demand is not rhetorical, and this year it arrives with more force than in any recent edition — for an arithmetical reason. The special report UNAIDS released at the opening records that international funding for the HIV response fell by roughly 25% from 2024 to 2025, from 8.3 to 6.2 billion dollars, with more than 1,700 service points closing. When external money contracts on that scale, the price of a technology stops being a negotiating problem and becomes a question of system sustainability. That shift is what reopens the intellectual property debate everywhere.

The demand is legitimate. It is also, in technical terms, imprecise — and the imprecision would not matter were it not for the mistaken expectation it creates about what happens after the decree is signed.

Brazil is arguably the country with the most consistent record in the world of using intellectual property flexibilities in the service of an access policy. That record is precisely what undermines the simple reading. What worked here was not an isolated act of force but an arrangement: a statutory duty to supply, a large-scale State purchaser, public manufacturing capacity and a credible threat. The compulsory licence was one component. Detached from the others, it delivers far less than is assumed.

One instrument and its neighbours

“Breaking a patent” is not a legal category. The instrument is a compulsory licence, and the difference has practical consequences. Under a licence the patent remains valid, the holder remains the holder and is remunerated; the State merely authorises another party to work the protected subject matter, temporarily and non-exclusively. The right is not extinguished — it is limited, for a price.

Brazil’s Industrial Property Law provides two distinct routes. Article 68 addresses abusive exercise of the right, abuse of economic power and failure to work the patent in national territory — the licence as a sanction for conduct. Article 71 addresses national or international emergency and public interest — the licence as an instrument of public policy. Different premises, different authorities, different timelines; conflating them weakens any claim.

Alongside them lie two routes that are often more effective and almost never discussed: invalidity, which asks whether the patent should ever have been granted, and forfeiture for failure to work. The distinction matters because the costs are asymmetrical. Invalidity attacks the title itself, has permanent effect and creates no diplomatic friction, but demands technical merit and time. A public-interest licence is quick on paper and expensive in foreign policy.

International law is not the obstacle

The notion that compulsory licensing offends the international order does not survive a reading of the texts. The Paris Convention of 1883 already admitted compulsory licences as a response to abuse of exclusive rights. Article 31 of the TRIPS Agreement regulates use without the holder’s authorisation by setting conditions — prior negotiation, waivable in cases of public non-commercial use, national emergency or extreme urgency; non-exclusivity; adequate remuneration reflecting the economic value of the authorisation; judicial review; and supply predominantly for the domestic market.

The 2001 Doha Declaration went beyond the text and settled the political ambiguity: each Member has the right to grant compulsory licences and the freedom to determine the grounds on which they are granted. Article 31bis, incorporated by the protocol in force since 23 January 2017, unlocked exports of medicines produced under licence to countries without manufacturing capacity — the so-called paragraph 6 system.

The conclusion is straightforward: access today is not blocked by a lack of international legal authorisation. It is a problem of institutional design and industrial policy.

The Brazilian record, read precisely

1996. Law 9,313 created the right to free antiretroviral treatment. Its legal effect is well known; its economic effect less so. A statutory duty to supply turns the State into a permanent, predictable purchaser, and it is that position — not rhetoric — that generates bargaining power. The statute itself, in fact, followed a court ruling that had already ordered the State to supply.

2001. Faced with the price of nelfinavir, the Ministry of Health announced a compulsory licence. The patent holder agreed to cut the price by around 40%, and the licence was never issued. The episode teaches something the current debate ignores: the instrument worked without being used. The value of a compulsory licence is not measured solely by the times it has been granted.

2000–2001. The United States took Brazil to the WTO dispute settlement system (DS199), challenging not the health-related licence but article 68 — the local working requirement. A panel was established on 1 February 2001 and the parties notified a mutually satisfactory solution on 5 July of the same year. Article 71, the basis for public-interest licences, was not at issue. The friction was never where common sense places it.

2007. Decree 6,108 of 4 May granted a public-interest compulsory licence over patents 1100250-6 and 9608839-7, covering efavirenz, for non-commercial public use within the National STD/Aids Programme, based on article 71 of the Industrial Property Law and article 4 of Decree 3,201/1999. Five years, renewable — and renewed in 2012. Remuneration of 1.5% of the cost of the medicine produced or the price of the product delivered to the Ministry of Health. A duty on the holder to disclose the information required to reproduce the protected subject matter. And, in article 5, express authorisation for the federal government to import the product should domestic manufacture prove unfeasible.

That last provision is the most revealing part of the decree. It is there because its drafters knew that authorisation and capability are different things. It remains the only compulsory licence for a medicine granted in the country’s history. In the nearly twenty years since, with obvious candidates — hepatitis C antivirals, high-cost oncology products, now lenacapavir — no other has been issued. That is not accident: it is the price of the instrument revealing itself.

What the licence does not deliver

Productive knowledge. A patent must describe the invention well enough for a skilled person to reproduce it. That falls well short of describing an industrial process: parameters, quality control, scale-up and accumulated tacit knowledge stay with the holder. For small molecules such as efavirenz, reverse engineering is feasible. For biologics — monoclonal antibodies, advanced therapies — the distance between being authorised to produce and being able to produce is vast: cell lines, process, comparability studies. A compulsory licence over a biologic without technology transfer risks being a piece of paper with no consequence.

Industrial base. A licence changes whom you buy from; it does not build where you produce. After 2007 Brazil turned to imported versions before manufacturing domestically. Without active pharmaceutical ingredients and qualified plants, one supplier is swapped for another — which may cut prices sharply, but does not create autonomy.

Regulatory passage. No decree replaces marketing approval. And here lies a policy incoherence that deserves more attention than it receives: Law 14,195/2021 repealed article 229-C of the Industrial Property Law, which had made the grant of pharmaceutical patents conditional on prior consent from the health regulator. The country removed a quality filter upstream, at the examination stage, in the same decade in which it expanded the discourse downstream, about licensing. Complaining about a monopoly after easing its constitution is a difficult position to hold.

Cost. Remuneration is owed. Law 14,200/2021 set 1.5% of net sales as a provisional floor until the amount is determined, taking into account the economic value of the licence, its duration, required investment, production costs and market price.

Credibility. A threat is an asset, and assets are spent. The country that issues compulsory licences is the same country that later negotiates prices, voluntary licences and production partnerships.

Two underused Brazilian advantages

The first is regulatory and almost invisible in the debate: Brazil grants no test-data exclusivity for medicines for human use. Law 10,603/2002 protects data submitted for the registration of veterinary products, fertilisers and agrochemicals — not human pharmaceuticals. In jurisdictions with data exclusivity, even an expired or licensed patent does not free the generic: registration remains blocked for years because the competitor cannot rely on the originator’s dossier. Here that blockage does not exist. It is a structural asset, under recurring pressure to change, and defending it may matter more in the medium term than any single decree.

The second is procedural: article 31 of the Industrial Property Law allows any interested party to file documents and information to inform the examination of a patent application, and invalidity remains available administratively and judicially. A significant share of the effective monopoly over medicines rests not on the molecule patent, often expired, but on secondary patents — formulation, polymorph, dosage, new use — of frequently debatable merit. Defeating them is slow, technical and quiet. It is also permanent, and costs no diplomatic capital.

What Law 14,200/2021 actually did

Anyone who studied this before 2021 must re-read article 71. The statute rewrote it entirely and added article 71-A. The current design is procedural: once a national or international emergency, a public interest or a nationwide state of calamity is declared, the executive publishes within 30 days a list of potentially useful patents, consults public bodies, teaching and research institutions and the productive sector, assesses each title individually within 30 days (renewable), and grants licences only to producers with proven technical and economic capacity.

Two provisions define the statute’s spirit. Paragraph 7 allows removal from the list of any holder who undertakes objective commitments capable of meeting domestic demand — direct exploitation, voluntary licensing or transparent supply contracts. Paragraph 18 requires the government, regardless of any licence, to prioritise technical cooperation agreements for acquiring the production technology and its transfer process. Read together, they show that the law did not make licensing easier: it formalised it and, in doing so, moved the instrument into the field of regulated bargaining. Article 71-A internalised the logic of article 31bis, allowing humanitarian licences for export to countries without manufacturing capacity.

The 2026 test

In January 2026 Brazil’s health regulator approved lenacapavir, administered twice a year, both for pre-exposure prophylaxis and for multidrug-resistant HIV. The product substantially reduces the risk of infection and changes the logic of prevention: two injections a year instead of a daily pill. Brazil took part in the clinical research and was left out of the voluntary licence granted to generic manufacturers for roughly 120 countries — as were Argentina, Peru and Mexico. Patents have been granted in national territory. The US list price exceeds 28,000 dollars per person per year, while independent estimates put production at scale at a few dozen dollars per person per year; the demand humanitarian medical organisations brought to the Rio conference is a ceiling of 40 dollars per person per year across low- and middle-income countries. Before any distribution through the public system, two steps remain: price setting by the federal price regulator and an incorporation assessment by the national technology committee.

UNAIDS Executive Director Winnie Byanyima captured the design at Riocentro when she asked how Latin America can still be left outside licences for life-saving medicines. The question is apt and uncomfortable: the exclusion does not stem from a lack of technical capability — the region ran the trials — but from an income classification that treats middle-income countries as a market rather than as a public health problem. It is that classification, more than the patent itself, that Brazilian foreign policy has reason to contest.

The question that organises the decision is not whether the country may license, but what it wants to obtain. Negotiation and compulsory licensing deliver price. Production partnerships and technology transfer deliver capacity, more slowly and with less visibility. The two outcomes rarely come from the same instrument, and Brazilian policy has alternated between them without choosing — which is why each new technology returns the discussion to its starting point.

A position

I favour keeping the instrument permanently available and free of taboo, and using it when negotiation fails. It is lawful, it is constitutional — the Constitution guarantees the inventor a temporary privilege in view of the social interest and the technological and economic development of the country, a clause that subordinates the right to a purpose — and it is compatible with the multilateral system.

I also hold that treating the compulsory licence as the solution, rather than as one component among others, reproduces the illusion that access is a legal problem. It is legal in part. In the remainder it is industrial, regulatory and fiscal.

There is an asymmetry the public debate rarely states: the cost of a compulsory licence is immediate and visible — pressure, litigation, diplomatic noise — while the benefit is deferred and diffuse. That asymmetry explains the distance between how often the instrument is demanded and how rarely it is granted, in Brazil and worldwide. Recognising it is the condition for using it well.

A final observation about the week now ending. An international conference is a moment of exposure, and exposure is precisely the input negotiation requires: no patent holder enjoys being the subject of ten thousand people gathered in the city where the product was tested. Whether the pressure built at Riocentro converts into price, into a licence or into nothing depends less on what was said in the sessions than on what the country does in the months ahead, once the conference has left and what remains are the regulators, the patent office and the budget. That is the phase, without an audience, in which access policy is usually decided.

A necessary note for readers

A compulsory licence is a government decision, not a patient’s claim. No judge grants one at an individual request in a medicines case. What is litigated individually is the supply of treatment through the public system or a private health plan — a different body of rules.

The firm is available for a responsible review of your situation, in English.

About the author

Ramon Martins Andrade (Brazilian Bar, OAB/RJ 188.374) holds a law degree from UFRJ (2011), with master’s degrees from Université Sorbonne Nouvelle and from the Brazilian Naval War College (EGN). He is currently a researcher in Human Rights and Health at ENSP/Fiocruz.

If your situation resembles the one described here, get in touch with the office.

Ramon Martins Andrade (OAB/RJ 188.374) is a lawyer who graduated from UFRJ in 2011, with master’s degrees from Université Sorbonne Nouvelle and from the Naval War College (EGN, Brazilian Navy). He is currently a researcher in Human Rights and Health at ENSP/Fiocruz.

This content is for information and education only. It does not constitute legal advice, advertising of results or any guarantee of outcome, under Rule 205/2021 of the Brazilian Bar Association (OAB).