Are fluorinated drugs PFAS?

A drug called lenacapavir, an anti-HIV medicine produced by Gilead Sciences, made news this summer after a Phase 3 clinical trial of 5,300 women in Uganda and South Africa found that it was 100% effective at preventing infection. Patients receive the treatment as a shot every 6 months—much more convenient than a daily pill. It is, by all accounts, a lifesaving molecule.

It could also be considered a member of the PFAS family, according to some definitions.

PFAS, short for per- and polyfluoroalkyl substances, are infamous “forever chemicals,” typically bearing long carbon chains completely bedecked with fluorine atoms. Thousands of these compounds are used in a vast range of products, including surfactants, nonstick and waterproof coatings, and firefighting foam. Though PFAS are highly effective in these applications, they have an unfortunate habit of persisting and accumulating in the environment. Some have been linked to negative health effects such as decreased immune responses, developmental problems, and cancer.

In 2021, the year before lenacapavir was approved by the US Food and Drug Administration, the Organisation for Economic Co-operation and Development (OECD), an international policy forum, broadened its definition of PFAS. The revised definition encompasses all “fluorinated substances that contain at least one fully fluorinated methyl or methylene carbon atom.” In other words, all it takes for a molecule to belong in the PFAS category is a single carbon atom bound to two or three fluorine atoms and not bound to any other halogen or to hydrogen anywhere in a molecule’s structure.

Fully fluorinated carbon atoms are much used and much beloved by drug designers. For example, lenacapavir has three such carbons, artfully nestled within a larger complex structure. The COVID-19 drug nirmatrelvir (a component in Paxlovid) also has one, as does the antidepressant fluoxetine (Prozac). In fact, there are hundreds of medicines whose structures include that particular type of functional group, including many on the World Health Organization’s essential medicines list.

The European Chemicals Agency put forward a proposal in 2023 that would effectively ban all PFAS fitting the OECD definition in Europe. If the proposal becomes law, it’s estimated to take effect sometime in 2027. In response, the European Federation of Pharmaceutical Industries and Associations (EFPIA) issued a statement in September 2023 warning that such a comprehensive restriction would make drug development and manufacturing in the EU nearly impossible.

“I think there’s quite a lot to happen in the near future,” says David O’Hagan, an organic chemist at the University of St Andrews who researches fluorine chemistry. But he thinks things will “settle down in due course” as chemists, companies, and regulators figure out how to draw the lines between harmful PFAS and helpful medicines.

The enchanted atom

Fluorine has been “to some extent considered to be sort of magic” in medicinal chemistry, says Nicholas Meanwell, a medicinal chemist who retired from Bristol Myers Squibb in 2022. It is highly electron withdrawing but also very small, so it can alter a molecule’s electronic properties and conformation without changing the overall size very much. The C–F bond is also remarkably strong.

“It’s possible [that] we became a little bit indiscriminate in our use of fluorine,” Meanwell says. According to some estimates, 20% of recent small-molecule drugs contain at least one fluorine atom. But those atoms are not used randomly, Meanwell says. Over the past 2 decades, chemists have learned a lot about fluorine’s properties and how to use it strategically to achieve specific properties.

Strategically placed fluorinated functional groups can be used to fine-tune a molecule’s conformation, membrane-crossing ability, potency, and more. Trifluoromethyl (CF₃) groups in particular—the crux of the discussion about PFAS—are a popular handle for tweaking a molecule’s lipophilicity and metabolic stability. Lenacapavir is a great example: its 10 fluorine atoms, including two CF₃ groups, are most likely the reason its protection lasts for 6 months, Meanwell says.

The EU proposal, which applies to any PFAS produced on metric-ton scale or greater, makes exceptions for existing drugs and agrochemicals that have gained approval through other regulatory processes. Fluorinated reagents, intermediates, and other components necessary for synthesizing those drugs and agrochemicals will likewise get a pass. But clinical candidates that haven’t yet gained approval may not, and that’s the part that chemists and companies seem most worried about.

“It’s not immediately obvious how you replace fluorine and get the same properties,” O’Hagan says. Swapping out a single functional group is relatively simple for some types of products but can completely change a drug molecule’s pharmacology. If drug designers were no longer allowed to use trifluoromethyl groups, a lot of drug development projects would basically be sent back to square one, Kirsty Reid, EPFIA’s director for science policy, says in an email.

Meanwell says, “I hope that the legislation will be enlightened and understanding of the needs of medicinal chemists.”

A perfluorocarbon by any other name

Wendy Heiger-Bernays and coworkers at the Boston University School of Public Health published a paper 2 years ago on how various definitions of PFAS would classify 360 organofluorine pharmaceuticals approved between 1954 and 2021 (iScience 2022, DOI: 10.1016/j.isci.2022.10402).

Only 5 of those compounds would have qualified as PFAS under the OECD’s original 2018 definition, which requires PFAS to contain three or more fully fluorinated carbons connected to each other, or two such carbons connected by an ether group. The same number would fall under the definition used by the US Environmental Protection Agency’s Office of Chemical Safety and Pollution Prevention, which similarly requires multiple perfluorinated carbons positioned adjacent to each other. Under the updated OECD definition, the number of drugs on the list that could qualify as PFAS jumps to 107. Most of those bear only a single CF₃group.

Heiger-Bernays says the paper was largely meant to illustrate the implications of different definitions of PFAS, not necessarily to suggest that drugs be regulated as PFAS.

It’s important to remember that these definitions serve a goal, Heiger-Bernays says. In this case, she adds, the goal of including a broad swath of organofluorine compounds is to limit the amount of nondegradable organofluorine into the environment, which is different “than arguing that . . . all carbon fluorine containing molecules are toxic.” The PFAS of the greatest concern toxicologically are not molecules with single fluorinated carbons but those with longer perfluorocarbon chains. But there are also tens of thousands of potential PFAS, and many of them are not well studied.

David Andrews of the Environmental Working Group, an activist organization, says that “there’s a strong, clear chemical reasoning for the broader [OECD] definition.” For one thing, he says, it’s easy to understand and communicate to regulators and the public. Does the molecule in question have a perfluorinated carbon? Congrats, it’s a perfluorocarbon. Defining PFAS broadly leaves fewer loopholes for companies to substitute a known bad actor for a new molecule that’s technically not banned but will probably be just as problematic.

The newer definition also addresses the problem of trifluoroacetic acid (TFA) as a persistent environmental pollutant. TFA also happens to be a common reagent in organic chemistry and peptide synthesis. Many—but not all—trifluoromethylated molecules have the ability to generate TFA as a degradation product in the environment.

“I think we have to get rid of the nonessential uses” of TFA and anything that could break down into TFA, says Hans Peter Heinrich Arp, a professor at the Norwegian University of Science and Technology and coordinator of ZeroPM, a research project dedicated to preventing and removing persistent pollutants in the EU.

While fluorinated refrigerants and blowing agents are by far the biggest known source of TFA in the environment, pesticides and pharmaceuticals are also notable contributors, according to a 2022 study by the German Environment Agency. The list of pharmaceutical sources of TFA identified by the assessment include anesthetic gases and six other drugs: bicalutamide, celecoxib, efavirenz, flecainide, fluoxetine, and sitagliptin.

Arp says regulatory exceptions for essential medicines and agrochemicals are sensible right now. In the long run, though, he’d like to see more research on their contributions to TFA pollution and a phaseout of known TFA-generating molecules.

For people and planet

Medicinal chemists have assays to attempt to predict how a molecule is likely to be metabolized in the human body. But what happens to drugs once they’re excreted from the body or otherwise disposed of isn’t often given much attention. And we are currently excreting a lot of fluorinated compounds, Meanwell says.

Pharmaceuticals lingering in the environment, where they can disrupt ecosystems and contribute to antimicrobial resistance, is a well-known problem. In the EU, environmental assessments have been required for drug authorization requests since 2006 but don’t factor into approval decisions for human medicines.

Klaus Kümmerer, who researches sustainable chemistry and pharmacy at Leuphana University, says questions about the overlap between PFAS and pharmaceuticals highlight the need to consider environmental persistence and potential transformation pathways early in the drug development process.

Not all fluorinated drugs are going to be environmentally persistent or release TFA. It depends on those drugs’ individual chemistries—the fluorinated functional groups they contain and how they’re attached. “We cannot say fluorine is bad in itself. We have to really understand where it’s a challenge,” Kümmerer says. Scientists need the data to understand how a molecule might behave in the environment and the discernment to design pharmaceuticals that balance human medical needs with the long-term health of the planet.

For instance, Kümmerer and his colleagues designed an analog of the antibiotic ciprofloxacin (the molecule contains a single fluorine atom, which the researchers did not alter) to improve its biodegradability (ACS Sus. Chem. Eng. 2021, DOI: 10.1021/acssuschemeng.1c02243). The researchers changed the molecule’s cyclopropyl functional group to a tetrahydrofuran ring to make the drug more easily hydrolyzable at low pH in the bladder and therefore less likely to contribute to antimicrobial resistance in the environment.

Change isn’t going to happen overnight, Kümmerer says, but he sees more chemists starting to incorporate biodegradation into drug design. “It’s not about banning existing drugs,” he says. The aim is “to have a better situation in 10 or 15 years.” That better situation may involve fewer trifluoromethylated compounds, but they probably can’t disappear entirely.

Kümmerer says that the pharma industry tends to be risk-averse and skeptical of change but that the right combination of economics, policy, and environmental data can spur people to innovate. Finding alternatives takes time and money, but it can also generate new patents and new market opportunities.

Drug development today involves more testing and compliance with more regulations than it did 50 years ago, Kümmerer says. But “we have more treatments for more diseases now than ever before. We are chemists; we can do this. We should be a bit more self-confident.”