Yesterday was the deadline.

Japan's Consumer Affairs Agency has stated that the revised labelling provisions and mandatory Good Manufacturing Practice standards under the Foods with Function Claims system apply to relevant products manufactured on or after 1 September 2026. The transition period, which began when the amended Food Labeling Standards were published on 23 August 2024, ended on 31 August. From this week, a notified Food with Function Claims in tablet, capsule or similar supplement form must be made under GMP, the manufacturing discipline the pharmaceutical industry has treated as table stakes for half a century.

The reason the deadline exists is a body count. In the spring of 2024, Kobayashi Pharmaceutical's beni-koji red yeast rice supplements, sold lawfully under the FFC system, were linked to a cascade of kidney injuries. By 29 May 2024 the running record stood at 1,614 medical cases, 284 hospitalisations and five reported deaths, with the company examining possible causal links in scores more. The products carried function claims. They had never been approved by anyone, because the FFC system does not approve anything. It accepts notifications.

A ministerial meeting on 31 May 2024 ordered the system rebuilt. What came back, in the Food Labeling Standards amendment of 23 August 2024, was a package any pharmaceutical regulatory affairs professional would recognise: mandatory collection and prompt reporting of suspected health-hazard information from September 2024, advance notification of up to 120 days for products containing novel ingredients from 1 April 2025, mandatory GMP from 1 September 2026, and an annual self-inspection report to the Consumer Affairs Agency.

What did not come back was an efficacy gate. A company can still put a function claim on a product in Japan without any authority ever reviewing whether the claim is true. That asymmetry, repeated with local variations across the region's five main frameworks, is the subject of this story. The longevity category's fastest-growing segments, muscle health and metabolic support, sit almost entirely inside these frameworks, and companies choose their framework the way tax planners choose jurisdictions. The choice is rational because the frameworks draw the food-drug line in five different places. This piece walks the line through each of them, then asks what the routing behaviour, the enforcement record and the actual evidence base mean for a pharmaceutical sponsor whose competition never had to prove anything.

FIVE FRAMEWORKS, FIVE PLACES TO DRAW THE LINE

Japan draws the line at paperwork. The FFC system, introduced in April 2015 under the Food Labeling Act as a growth-strategy measure, lets a company label a food with a function claim after notifying the Secretary-General of the Consumer Affairs Agency, with the supporting evidence, at least sixty days before sale. The evidence can be the company's own clinical study or a systematic review of published literature; the agency checks completeness, not validity. Notification is not approval, and every label must say so: FFC products carry a mandatory statement that the product has not been individually evaluated by the government, alongside the notification number that lets anyone retrieve the filed claim and evidence from the CAA's public database. By 22 March 2024, on the eve of the beni-koji crisis, 1,693 companies had notified 6,795 products, and the research firm Fuji Keizai put the 2023 market at 686.5 billion yen, roughly twenty-fold its size in the system's first year. The claim vocabulary is effectively open within the boundary of not naming disease. Muscle is a settled category: notified products carry claims on maintaining muscle mass and supporting muscle strength in middle-aged and older adults, filed on ingredient evidence for compounds such as beta-hydroxy-beta-methylbutyrate, and the claim text is public because the register is.

China draws the line at the gate. Health food, the blue-hat category, has required pre-market registration or filing since the dual-track system under the Administrative Measures for Health Food Registration and Filing took effect on 1 July 2016. Filing is the light lane, reserved for nutrient supplements whose ingredients and claims match the official catalogues. Everything else, which means every functional product with pharmacological ambition, goes through registration: state-reviewed toxicology, functional testing and technical evaluation, on timelines practitioners describe in years. The claims themselves come from a closed list. The Directory of Health Functions Available to be Claimed by Health Food (Non-nutrient Supplements) (2023 Version), released in August 2023, permits twenty-four function claims, and a product may carry only the claim it was registered for, phrased the way the directory phrases it. Every label must also carry the warning, mandatory since the SAMR health food labelling guidance took effect on 1 January 2020, that health food is not medicine and cannot substitute for medicine in treating disease. The wall is real and the traffic through it is thin: SAMR data compiled at 24 December 2024 showed 392 domestic registration certificates and 4,210 filings granted in 2024, and between 2018 and 2024 not a single imported health food obtained registration. That last number explains most of what follows in this story, because a wall that high does not stop demand. It reroutes it.

Korea draws the line at the ingredient. The Health Functional Food Act, first promulgated on 27 August 2003 and administered by the Ministry of Food and Drug Safety, splits the world into notified functional ingredients, listed in the Health Functional Food Code and usable by anyone, and individually recognized functional ingredients, a system running since 2004 in which a company submits safety and functionality data and, if MFDS grants recognition, obtains an ingredient claim it alone may use. A peer-reviewed analysis published in February 2026 counted forty-one distinct functional categories recognized by 2025, and the newer additions read like a longevity clinic's menu: muscle strength maintenance, sleep quality, stress regulation, menopausal care. MFDS recognition data show where industry is heading; in 2022, muscle strength improvement applications were rising alongside memory and cognitive function. Around the claims sits more machinery than any peer framework: pre-review of advertisements, a duty on business operators to report adverse events, and re-evaluation of functional ingredients every ten years. The market this supports reached roughly 4.18 billion US dollars in 2024 on MFDS production data. Korea's is the most pharmaceutical of the five architectures in structure, with one deliberate omission: the recognition standard is functionality evidence graded for food, not therapeutic efficacy, and the exclusivity it grants is a commercial instrument, not a licence.

India draws the line in pencil. The category exists under Section 22 of the Food Safety and Standards Act 2006, and was first given standards by the FSS (Health Supplements, Nutraceuticals, Food for Special Dietary Use, Food for Special Medical Purpose, and Prebiotic and Probiotic Food) Regulations 2016. Their intended replacement, the draft 2022 regulations, has never been finally notified. Instead FSSAI has operationalized the draft by direction, first from 29 March 2022 and 10 May 2022, re-operationalized from 1 October 2022, and re-operationalized again with effect from 1 July 2024 under Section 16(5) of the Act, each direction acknowledging that finalisation will take more time. The rules that do bind are ingredient rules: botanicals must trace to the Indian Pharmacopoeia, the Ayurvedic Pharmacopoeia or other recognised compendia under Regulation 16, and vitamins and minerals are capped by Schedule IV, generally at recommended dietary allowance levels. What India does not have is product-level claim clearance. No authority reviews a nutraceutical's claim before it is made; the FSS (Advertising and Claims) Regulations 2018 police it afterwards, if at all. The government knows. An inter-ministerial committee chaired by former health secretary Apurva Chandra, constituted in early 2024 with the heads of both FSSAI and the drugs regulator CDSCO among its eight members, reported in November 2024 that claims to cure, mitigate or reduce the risk of disease should move to CDSCO, with FSSAI retaining only the nutritional and health claims listed in the 2018 regulations, and dose-format vitamin, mineral and amino acid products with pharmaceutical excipients shifting toward drug rules. As of April 2026 the recommendations remained recommendations, with public representations still urging the health ministry to implement them.

Australia draws the line at the dictionary. A listed medicine, marked AUST L, enters the Australian Register of Therapeutic Goods under section 26A of the Therapeutic Goods Act 1989 on the sponsor's self-certification. The TGA does not check safety, quality or efficacy before sale, a fact the regulator itself states plainly. What Australia controls instead is vocabulary: every ingredient must appear in the Permissible Ingredients Determination within its specified limits, and every indication must be selected from the permitted indications list, a closed set of roughly 1,300 low-level claims. A sponsor who wants an efficacy-assessed claim can use the AUST L(A) pathway, in which the TGA reviews the evidence behind the indication before listing; the Evidence Guidelines were reissued as Version 4.0 on 20 November 2024 to tell sponsors what they must hold. The structure is the inverse of Japan's: entry is trivially easy, but the claim language is the most tightly rationed in the region. The check arrives later, through post-market compliance reviews, and its findings are the least flattering public record any of the five regulators maintains about its own category. We return to it below.

WHERE THE CATEGORIES OVERLAP, AND HOW ROUTING GETS DECIDED

Put the five side by side and the same product concept can be a registered quasi-pharmaceutical object in China, an exclusive recognized ingredient in Korea, a notified food in Japan, a self-certified listed medicine in Australia and an essentially unexamined nutraceutical in India. Companies read that table the way capital reads tax treaties, and no molecule illustrates the routing better than nicotinamide mononucleotide, the NAD precursor that has become the longevity retail category's signature ingredient.

In China, NMN is prohibited in food. A SAMR document of 13 January 2023 directed local authorities to investigate the sale and marketing of so-called anti-aging products built on NMN, clarifying that the compound has been approved as neither food additive, new food raw material, drug ingredient nor health food ingredient. The only lawful channel is cross-border e-commerce, where NMN rides the positive list as a vitamin B derivative and function claims are prohibited; in 2024, platforms further restricted which origin countries' NMN products could be sold. In Japan, over the same period, NMN has been accepted for use in notified Foods with Function Claims. In Australia, the TGA issued an alert on 7 April 2025 stating that NAD and NMN were not permitted ingredients in listed medicines and that enforcement action could follow their appearance in product names or advertising, and then, on 10 December 2025, added NMN to the Permissible Ingredients Determination (No. 4) 2025 on an application by SyncoZymes, granting the applicant a two-year exclusivity on the ingredient's use in listed medicines to 10 December 2027. One molecule, three regulatory identities, and in one market both identities inside a single calendar year.

The routing logic this teaches is simple. A wall like China's does not remove a market; it moves the point of sale offshore and the claim into unregulated marketing. An open notification system like Japan's becomes the regional landing zone for functional ambition, which is why its safety failure mattered so much. And a vocabulary system like Australia's pushes the pharmacology into whatever phrasing the permitted list will bear, which is why the TGA's compliance reviews keep finding listed medicines advertising outside their sanctioned wording. Sponsors do not choose the strictest defensible framework. They choose the most permissive available one, and the muscle health and metabolic categories are growing fastest precisely where the gate is lightest: India's protein and supplement boom is running on a framework that has never finally notified its own regulations, prompting the Indian Council of Medical Research's 2024 dietary guidelines to advise the public against protein supplements for muscle building even as the market compounds.

THE ENFORCEMENT RECORD

Read regulator by regulator, the last thirty months are the story of a category being disciplined on everything except its claims to work.

Japan's record is the reform itself: hazard-reporting duties in force from September 2024, the 120-day novel-ingredient window from April 2025, GMP binding on products manufactured from 1 September 2026, annual self-inspection reports, and framed FFC identification on the front of pack. Korea's record is vigilance at the marketing edge: an MFDS inspection of online advertising between 28 October and 12 December 2025 found sixteen companies in breach of the Act on Labeling and Advertising of Food, including AI-generated videos of fake physicians attributing implausible effects to NMN products, with results announced on 15 December 2025 and expanded surveillance of synthetic expert endorsements announced the following day; in July 2026, MFDS moved to further strengthen GMP requirements for functional foods by legislative proposal. China's record is the wall doing its work: the January 2023 NMN directive, platform-level sales restrictions, and a registration channel so demanding that imports have simply stopped attempting it. India's record is the gap between diagnosis and treatment: a committee co-signed by the heads of both its food and drug regulators recommending in November 2024 that disease-adjacent claims leave the food framework, and a food framework still operating by renewable direction while the recommendation waits.

Australia's record deserves its own paragraph, because it is the region's most transparent audit of what self-certification produces. The TGA publishes its compliance review outcomes in a public database. The March 2025 update flagged listed medicines making non-permitted claims about ADHD, asthma, cardiovascular disease, diabetes, urinary frequency and, notably for this series, exercise performance. The March 2026 update covered thirty-two listed medicines and included the cancellation and recall of nineteen sunscreens that shared a base formulation with Ultra Violette Lean Screen SPF 50+, formerly AUST L 332788, after independent testing published by CHOICE in 2025 found lower-than-claimed sun protection; a further nine products in the same update were cancelled from the register and two recalled. Historical TGA review programmes have reported non-compliance in half or more of reviewed products, a rate the regulator's risk-targeting partly explains and only partly excuses. The honest reading is that the AUST L number on a longevity product certifies a promise, not an inspection, and the TGA's Compliance Principles for 2026 and 2027 signal that the regulator knows the gap it is policing.

Set the records side by side and the convergence is unmistakable. GMP is now mandatory or hardening in Japan, Korea, China and the Indian committee's recommendations. Adverse event collection is a legal duty in Japan and Korea. Korea re-evaluates its ingredients on a ten-year cycle the way drug regulators revisit approvals. Exclusivity, the classic pharmaceutical incentive, now exists inside food and complementary frameworks in Seoul and Canberra. Piece by piece, the region's food regulators are borrowing the pharmaceutical rulebook. The one chapter nobody has borrowed is the one that made the rulebook worth having: prove the product does what the label implies before you sell it.

COMPETING AGAINST A PRODUCT THAT NEVER HAD TO PROVE ANYTHING

For a pharmaceutical sponsor developing in the aging-adjacent space, this landscape is not background noise. It is the competitive set, and it holds three lessons.

The first is the asymmetry of time and cost. A muscle-health claim can reach a Japanese shelf sixty days after notification and an Indian shelf with no claim review at all, while a sponsor pursuing sarcopenia as an indication faces the endpoint politics this series mapped in LG-03 and, before that, the absence of an approvable aging indication mapped in LG-01. The supplement claim and the drug indication are not different grades of the same permission. They are different products of different systems, separated by roughly five orders of magnitude in evidence cost, and they land on the same consumer with similar sentences. Any commercial model for a gerotherapeutic that assumes a claims-free competitive field is wrong on day one, and the pricing floor problem LG-06 and LG-07 described gets harder when the reference product next to yours retails for the price of groceries.

The second is that claim vocabularies are endpoint definitions by other means. When China's 2023 directory added bone and joint maintenance to its permitted list, when MFDS recognized muscle strength maintenance as a functional category, when Australia's permitted indications set the exact sanctioned phrasing for muscle claims, each regulator was writing a definition of what counts as demonstrated muscle benefit, on evidence standards far below AWGS 2025 or GLIS, and each definition now shapes what consumers in that market believe the words mean. A sponsor who eventually wins a sarcopenia indication will market it into a population already saturated with adjacent language that cost its users almost nothing. The endpoint is the asset, and in five markets the asset has already been minted in a softer metal.

The third is that the tightening wave is an opening. GMP mandates, hazard reporting and advertising surveillance raise the cost base of the supplement channel and thin out its least serious operators; Japan's transition just did this in real time. Sponsors with pharmaceutical-grade manufacturing and vigilance already sunk into their cost structure lose least from this convergence, and the emergence of assessed pathways, Australia's AUST L(A) and Korea's individually recognized ingredients, gives evidence-rich players a way to convert data into claims competitors cannot copy. The strategic question is no longer whether to take the food channel seriously. It is whether to hold the disease claim as the destination while the food frameworks, rung by rung, price everyone else up toward pharmaceutical discipline.

THE EVIDENCE THAT DOES EXIST

None of this means the shelf is empty of science, and in muscle health and protein nutrition the record deserves to be stated straight, because it is the strongest in the longevity retail category.

Expert consensus groups have for a decade recommended protein intakes for older adults above the general adult allowance, in the range of 1.0 to 1.2 grams per kilogram per day, on evidence that habitual intakes in aging populations undershoot the requirements of muscle protein maintenance. Interventional data exist at registrational scale for the category's standards: the PROVIDE trial, published in 2015, randomized 380 sarcopenic older adults to thirteen weeks of a vitamin D and leucine-enriched whey protein supplement or control and reported improvements in appendicular muscle mass and in chair-stand performance. For beta-hydroxy-beta-methylbutyrate, the leucine metabolite that anchors many notified muscle claims in Japan, a 2025 meta-analysis in Frontiers in Nutrition, from authors including researchers at China's National Center for Food Safety Risk Assessment, pooled twenty-one randomized controlled trials and concluded that supplementation at three grams daily for longer than twelve weeks measurably benefits muscle parameters that decline with age. Creatine, combined with resistance training, has meta-analytic support for lean mass gains in older adults that most drug programmes in the space would be pleased to match. And the exercise co-dependency running through all of it is not a weakness of the evidence; it is the evidence, consistent with everything LG-03 reported about how muscle endpoints move.

The honesty the category cannot supply is proportion. These are modest effects, in heterogeneous trials, on endpoints chosen by investigators rather than negotiated with regulators, and they attach to commodity ingredients: protein, creatine, a leucine metabolite. Here is the perverse close of the loop. The best-evidenced substances in the longevity retail category are the ones nobody can own, so the claim systems' newest instruments, Korea's exclusivity and Australia's, reward the ownable novel ingredient over the evidenced commodity. The frameworks are not merely failing to require proof. At the margin, they are pricing it against itself.

Five frameworks drew the food-drug line in five places, and for a decade the products moved to wherever the line was kindest. The regulators' answer, visible from Tokyo's GMP deadline this week to Seoul's advertising sweeps to Delhi's unimplemented committee report, has been to move the manufacturing, the vigilance and the paperwork across the line to the pharmaceutical side. The proof stays where it always was. Until a framework moves that too, the most consequential sentence in Asia-Pacific longevity retail remains the one printed on every blue-hat label in China: this is not medicine. The label is right. The market keeps buying the ambiguity anyway.

(arcilla.fran@biopharmaapac.com)