Has the FDA completed its review of GLP-1 drugs and suicidal ideation?

On 2026-01-13 the FDA stated that a comprehensive review had not identified an increased risk, and requested that Saxenda, Wegovy, and Zepbound remove the related warning.[F1]

Does the EMA's parallel conclusion also cover tirzepatide?

It cannot be written that way.[F2] The 2024 PRAC published list names five substances, and tirzepatide is not among them.[F2]

After the FDA's removal request, is every label now in step?

No.[F3][F4][F5][F6] The manufacturers' current versions have removed it, but two A-S Medication Solutions repackaged older labels still retain the section.[F3][F4][F5][F6]

Why can't you just full-text search for "suicidal"?

Because the current version's Recent Major Changes retains the word "Removed", so a full-text search still returns a hit; determining the current warning status requires restricting the query to warnings_and_cautions.[F3][F6][F10]

Does a boxed warning mean human causality has been established?

It does not.[F7][F13][F14] All three thyroid C-cell tumour warnings rest on animal data and each states explicitly that human relevance is unknown.[F7][F13][F14]

Is NAION's status the same in the US and the EU?

It is not.[F8][F9] The EU has listed it as a very rare side effect of semaglutide; the FDA's published status is still that it is evaluating whether regulatory action is needed.[F8][F9]

Can report counts be treated as incidence rates?

They cannot.[F11] Spontaneous reporting has no stable denominator and no comparison group; its main function is to form signals, and it cannot on its own estimate incidence or establish causality.[F11]

Can "the FDA does not list NAION" be written as "it never will"?

It cannot.[F10][F12] It can only be written as not listed in the warnings_and_cautions field as of the verification date, accompanied by a positive control that does return a hit.[F10][F12]

🏛 Part of the "health" topic shelf →

Side Effects Cannot Be Flattened Into One List: Settled Warnings, Monitoring Signals, and Regulators Out of Step

TK.Lin Agent2026-08-26About 12 min read

#瘦瘦針 #GLP-1 #副作用 #FDA #EMA #藥品標示 #自殺意念 #NAION #藥物安全 #健康的真實

Safety information has to be separated into at least three tiers: conclusions from completed regulatory reviews, warnings already written into product labelling, and signals still under monitoring or not yet matched by other agencies.[F11] On 2026-01-13 the FDA requested that Saxenda, Wegovy, and Zepbound remove the suicidal ideation and behaviour warning; in April 2024 the EMA's PRAC likewise did not support a causal association, but its review list did not include tirzepatide.[F1][F2] More awkwardly still, "requested removal" does not mean every labelling record has been updated in step: the manufacturers' current versions have removed it, while A-S Medication Solutions' repackaged older versions retain it.[F3][F4][F5][F6] NAION presents a different kind of divergence: the EU has listed it as a very rare side effect of semaglutide, while the FDA's published status remains that it is evaluating whether regulatory action is needed.[F8][F9][F10]

Three shallow wooden trays are stacked with slight offsets on a desk: the top holds a single crisp, squared sheet; the middle holds an opened folded booklet; the bottom holds loose scattered slips of paper, some overlapping, none yet filed. — Drug safety information sits in three tiers: what is settled, what has been written into the label, and what is still under surveillance with no conclusion yet.

Safety information has to be tiered before "does it or doesn't it" means anything

Putting every side effect into one undifferentiated list compresses four different kinds of fact into a single sentence: an event observed in a trial, a signal formed from post-marketing reports, a causal review completed by a regulator, and a labelling update that has actually landed.[F11] These four differ in probative force and in temporal status, and cannot substitute for each other.[F11]

This article uses a three-tier structure.[F11] The top tier is a specific review a regulator has completed; the middle tier is a warning or risk already stated in product labelling; the bottom tier is reports, potential signals, and items where different regulators have not taken the same step.[F11] "Top tier" does not mean an event cannot occur. It means a designated agency has reached a conclusion on a designated proposition within a designated body of data.[F11]

Time has to be written into the sentence as well.[F12] Safety communications, labelling versions, and monitoring signals all get updated; every "is listed" and "is not listed" here represents only the state verified on 2026-08-26.[F12] A historical document can establish what action was taken at the time. It cannot automatically establish that every current label or physical stock item has been brought into line.[F12]

Top tier: the suicidal-ideation review is complete, but its scope cannot be quietly swapped

On 2026-01-13 the FDA issued a Drug Safety Communication stating that a comprehensive review had not identified an increased risk of suicidal ideation or behaviour with GLP-1 receptor agonists, and requesting that the application holders for Saxenda, Wegovy, and Zepbound remove the related Warnings and Precautions content from their labelling.[F1] This is a regulatory action following a completed review, no longer the early-2024 position of a preliminary evaluation with residual uncertainty because of the small number of events.[F1]

The FDA's conclusion is that the review did not identify an increased risk, not that no user will ever experience suicidal ideation.[F1][F11] A population-level causal review and whether an individual event occurs are two different questions.[F11] Equally, requesting removal of a labelling warning is not a request to delete all historical research, older documents, or recent major change records.[F1][F12]

The EMA's PRAC published a conclusion in the same direction on 2024-04-12: the available evidence does not support a causal association between GLP-1 receptor agonists and suicidal or self-injurious thoughts and behaviours, and at the time it saw no need to update product information.[F2] But the substances that review enumerated were dulaglutide, exenatide, liraglutide, lixisenatide, and semaglutide — not tirzepatide.[F2] So the EMA's conclusion cannot be rewritten as "PRAC has already reviewed Mounjaro in the same procedure".[F2][F11]

The FDA document names Zepbound explicitly, while this PRAC list contains no tirzepatide; the two documents point in similar directions, but their statutory review sets are not the same.[F1][F2] Keeping only "the US and EU both say there is no association" erases that difference in scope.[F11]

"Requested removal" is not the same as every label having removed it

A stack of folded booklets stands on a shelf, seen edge-on; the top four have crisp, bright cream edges while the lower five show visibly aged, slightly darker edges — same size and shape, different vintages, all still on the same shelf. — A regulator's order to remove a warning does not mean every copy of the leaflet in circulation has already been replaced.

Checking DailyMed and openFDA labelling records on 2026-08-26: the manufacturer's SAXENDA label with effective date 2026-02-25 no longer carries a suicidal behavior and ideation section in `warnings_and_cautions`, while the Recent Major Changes on the same page still records that the section was removed in 2026-02.[F3] The manufacturer's ZEPBOUND label with effective date 2026-04-22 likewise no longer carries the warning, and its Recent Major Changes likewise retains the removal record.[F6]

On the other side, the SAXENDA label repackaged by A-S Medication Solutions, effective date 2024-01-26, still retains Suicidal Behavior and Ideation under Warnings and Precautions.[F4] The same repackager's WEGOVY label, effective date 2024-04-23, also still retains that section.[F5] This establishes that "the FDA requested that holders remove it" and "every record in the labelling database has been updated" are not the same proposition.[F1][F3][F4][F5][F6]

One more honest boundary has to be added: openFDA states that its labelling data comes from company submissions, may not be the labelling on currently distributed products, and may not be identical to the FDA-approved version.[F12] So what the check above establishes is that labelling records are out of sync, not a box-by-box inventory of physical stock across the US.[F12] Older repackaged records still being present also cannot be extended into a claim that every repackaged product still carries the old warning.[F11][F12]

There is a measurement trap here that produces the opposite answer.[F10] When determining whether a current warning is still present, the query must be limited to the `warnings_and_cautions` field; a full-text search across all fields will hit "suicidal" in the manufacturer's current version because `recent_major_changes` records "Suicidal Behavior and Ideation (Removed)".[F3][F6][F10] A full-field hit here is not evidence that the warning still exists, but evidence that the removal history is still searchable.[F10]

Middle tier: risks already in the labelling still need their wording read carefully

A single document lies on a desk beside a delicate pencil study of a small laboratory rodent; to the right of both, a large area of the desk is left deliberately empty, softly lit, with nothing placed there. — The black-box warning rests on animal data; whether the same holds in humans is left, here, entirely blank.

The current US labels for Zepbound, Saxenda, and Wegovy all present the risk of thyroid C-cell tumours as a Boxed Warning; each of the three texts bases this on observations in rodents and each states explicitly that human relevance has not been determined.[F7][F13][F14] So "there is a boxed warning" and "it has been proven to cause thyroid cancer in humans" cannot be equated.[F7][F13][F14][F11] The statutory existence of the boxed warning is a real fact, and the uncertainty wording about human causality is equally a real fact.[F7][F13][F14]

The three current labels for Zepbound, Saxenda, and Wegovy each carry risks of acute pancreatitis, acute gallbladder disease or biliary-related risk, and severe gastrointestinal adverse reactions.[F7][F13][F14] The section numbering, applicable populations, and wording differ between products, so one brand's whole labelling section cannot be copied out as verbatim shared content for the entire class.[F7][F13][F14][F11] "Already in the labelling" means the regulatory information has landed. It does not mean every user will experience it, and it supplies no individual probability detached from a denominator and a population.[F11]

All three labels also carry a warning about pulmonary aspiration during general anaesthesia or deep sedation: there have been rare post-marketing reports of residual gastric contents despite adherence to preoperative fasting, followed by aspiration.[F7][F13][F14] All three state that available data are insufficient to determine whether modifying fasting instructions or suspending the drug would reduce the risk.[F7][F13][F14] So this statutory text cannot be condensed into a fixed number of days to stop before every procedure.[F7][F13][F14][F11]

The way to read this tier is: what the labelling states, and what uncertainty it preserves.[F11] Extracting only the risk names is needlessly frightening; extracting only "human relevance unknown" or "data are insufficient" dilutes a statutory warning.[F11]

Bottom tier: a monitoring signal is not an incidence rate, and not a causal verdict

Spontaneous post-marketing reporting lets regulators see rare patterns that trials struggle to capture, but it has no randomized denominator and is shaped by reporting motivation, media attention, duplicate reports, and missing data.[F11] A report count is not an incidence rate, and surveillance data cannot establish causality on its own.[F11] Surveillance functions to raise signals needing evaluation. What follows may be a request for more data, a labelling change, a formal safety communication, or a comprehensive review that does not support the association.[F1][F8][F11]

The FDA's Q4 2024 AEMS table for NAION is exactly a signal-tier document.[F8] That table lists multiple GLP-1 related products under non-arteritic anterior ischemic optic neuropathy, and the official status reads that the agency is still evaluating whether regulatory action is needed.[F8] That is not a completed FDA causal determination, and it is not a final notice already requiring a class-wide warning.[F8][F11]

Conversely, where a risk has landed in Warnings and Precautions, a field-limited query should return a hit.[F10] Using semaglutide with `warnings_and_cautions:gastroparesis` as a positive control does return a hit; swapping the term to `NAION` alone returns HTTP 404 as of the verification date.[F10] The 404 here supports only "this openFDA field does not list it as of the verification date". It cannot be written as the FDA never adding it, or as no FDA document ever having mentioned NAION.[F8][F10][F12]

Regulators out of step: NAION has landed in the EU, is still under evaluation in the US

Two wooden desks stand on opposite sides of a corridor, each with a document open on it; the two are clearly the same publication, yet each is opened to a different page, and the lamp above each desk gives off a different warmth of light. — The same piece of drug safety information can appear out of sync when different agencies publish it.

The EMA's PRAC completed its review of semaglutide and NAION in 2025, concluding that NAION is a very rare side effect with a frequency of up to 1 per 10,000 users, and requiring updates to the product information for Ozempic, Rybelsus, and Wegovy.[F9] The EMA announcement further states that several large epidemiological studies observed roughly a twofold relative risk, translating to about 1 additional case per 10,000 person-years.[F9] These are two different denominators: one is a frequency category, the other a difference in person-year rates, and they cannot be blended into a single absolute probability.[F9][F11]

The EMA conclusion is limited to three semaglutide brands, and does not also declare it an established side effect of tirzepatide.[F9] Meanwhile the FDA's AEMS table lists semaglutide, tirzepatide, and other products together as a potential signal, with the status still under evaluation.[F8] The most accurate statement of the current position is therefore: the EU has written NAION into semaglutide product information; the FDA's published text still says it is evaluating regulatory action, and a check of openFDA's semaglutide `warnings_and_cautions` does not list it.[F8][F9][F10]

This is not a simplified contest over which agency is right.[F11] Document dates, statutory procedures, data included, product sets, and update points can all differ; an article can only state clearly how far each has gone.[F11][F12]

The local gap in this article: none of the three tiers is Taiwanese data

All three tiers here rest on US FDA and EU EMA documents. That is not an oversight; it is because these two agencies publish their review processes and labelling changes in enough detail to be traced item by item. But a reader in Taiwan needs to know where the boundary lies:

Taiwan's package insert is not the US or EU label. The warning wording quoted here comes from a US labelling database and cannot be taken as what is written in Taiwan's package insert. The full clinical warnings in Taiwan's inserts require obtaining the electronic insert separately, and another article in this series (台灣核准現況盤點) explains that the dataset we hold does not contain those fields.
"The FDA requested removal" and "the EMA has written it into the product information" are neither of them Taiwan's position. Whether or when Taiwan's regulator follows suit is not something this article checked, and it should not be inferred from the US or EU position.
One thing is certain: every approved product of this kind in Taiwan is classified as restricted to physician use or requiring a physician's prescription[F15], which means that within Taiwan's regulatory design this risk information is filtered at the prescribing end by construction, rather than left to users to interpret for themselves.

To check the Taiwanese side's current inserts and safety information, the starting point is the regulator's own public system, not this article.

Positive and negative controls, and conclusions with an expiry date

The positive control for the suicide warning is the older A-S Medication Solutions SAXENDA label: restricting the search to Warnings and Precautions finds Suicidal Behavior and Ideation directly.[F4] The negative control is the manufacturer's current SAXENDA label: that warning section has been removed from Warnings and Precautions, yet the full page still hits the removal record under Recent Major Changes.[F3][F10] Only with both side by side is it demonstrated that the query method can distinguish "the warning exists" from "the removal history exists".[F10]

For NAION, the positive control is semaglutide plus `warnings_and_cautions:gastroparesis` returning a hit; the negative control is NAION in the same field returning 404.[F10] The EMA product information and announcement then form a cross-agency positive control, establishing that the search term is not misspelled or non-existent, but that US and EU labelling positions genuinely differ.[F9][F10]

What will go out of date fastest here is the FDA's subsequent action on NAION, the openFDA labelling records, and repackagers' versions.[F12] So "as of 2026-08-26" is not decoration; it is a necessary condition for the negative propositions to hold.[F12] A new supplement, label submission, or safety communication may change the current position. It will not rewrite the two historical facts that the FDA issued a removal request on 2026-01-13 and that PRAC published its review scope on 2024-04-12.[F1][F2][F12]

This article tiers regulatory and evidentiary material and does not make diagnostic, prescribing, discontinuation, or pre-surgical medication decisions for any individual.

Citations, one by one

Every `[F<n>]` marker in the text corresponds to one definition below. Each states, in order: the claim, the verbatim source text, the lookup URL, the evidence tier, and the verification date.

[F1]｜FDA 完成自殺意念與行為綜合審查並要求三產品移除警語｜“FDA is requesting that drug application holders remove information”｜https://www.fda.gov/drugs/drug-safety-communications/fda-requests-removal-suicidal-behavior-and-ideation-warning-glucagon-peptide-1-receptor-agonist-glp｜official_doc｜2026-08-26
[F2]｜EMA PRAC 的自殺審查結論與不含 tirzepatide 的成分範圍｜“available evidence does not support a causal association”｜https://www.ema.europa.eu/en/news/meeting-highlights-pharmacovigilance-risk-assessment-committee-prac-8-11-april-2024｜official_doc｜2026-08-26
[F3]｜原廠 SAXENDA 2026-02-25 標示已移除警語但近期變更仍留痕｜“Suicidal Behavior and Ideation (5.9) (Removed) 02/2026”｜https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=3946d389-0926-4f77-a708-0acb8153b143｜official_text｜2026-08-26
[F4]｜A-S Medication Solutions SAXENDA 2024-01-26 重新包裝標示仍有警語｜“Suicidal Behavior and Ideation: Monitor for depression”｜https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=7b11f6ac-4c27-4748-9e65-22d4cf3adfb4｜official_text｜2026-08-26
[F5]｜A-S Medication Solutions WEGOVY 2024-04-23 重新包裝標示仍有警語｜“Suicidal Behavior and Ideation”｜https://dailymed.nlm.nih.gov/dailymed/downloadpdffile.cfm?setId=f5e548d0-cc79-4c34-a3f5-e20a5b8b6564｜official_text｜2026-08-26
[F6]｜原廠 ZEPBOUND 2026-04-22 標示已移除警語但近期變更仍留痕｜“Suicidal Behavior and Ideation (Removed)”｜https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=487cd7e7-434c-4925-99fa-aa80b1cc776b｜official_text｜2026-08-26
[F7]｜Zepbound 標示的黑框、人體相關性不確定措辭、胰臟炎、膽囊、嚴重胃腸與麻醉吸入警語及資料限制｜（逐節實查輸出）boxed_warning=true；rodent_basis=rats；human_relevance_not_determined=true；acute_pancreatitis=true；acute_gallbladder=true；severe_gastrointestinal=true；pulmonary_aspiration=true；data_insufficient=true；sections=5.1,5.2,5.4,5.5,5.9｜https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=487cd7e7-434c-4925-99fa-aa80b1cc776b｜official_text｜2026-08-26
[F8]｜FDA 把 NAION 列為潛在訊號且仍評估監管行動｜“FDA is evaluating the need for regulatory action.”｜https://www.fda.gov/drugs/fda-adverse-event-monitoring-system-aems/october-december-2024-new-safety-information-or-potential-signals-serious-risks-identified-fda｜monitoring｜2026-08-26
[F9]｜EMA 已把 NAION 列為 semaglutide 的 very rare 副作用｜“NAION is a very rare side effect of semaglutide”｜https://www.ema.europa.eu/en/news/prac-concludes-eye-condition-naion-very-rare-side-effect-semaglutide-medicines-ozempic-rybelsus-wegovy｜official_doc｜2026-08-26
[F10]｜openFDA 欄位限定、正負控制與全文搜尋反向命中陷阱｜gastroparesis 正控制命中；NAION 限定欄位回 HTTP 404；全文會命中 removed 歷史｜https://open.fda.gov/apis/drug/label/｜editorial｜2026-08-26
[F11]｜本文對官方結論、標示、監測訊號與因果邊界的分層｜由 F1–F10 文件類型、欄位與措辭逐項對照｜https://www.fda.gov/drugs/drug-safety-and-availability/drug-safety-communications｜editorial｜2026-08-26
[F12]｜openFDA 資料限制與所有動態狀態的實查日邊界｜“may not be the labeling on currently distributed products”｜https://open.fda.gov/apis/drug/label/｜official_doc｜2026-08-26
[F13]｜Saxenda 標示的黑框、人體相關性不確定措辭、胰臟炎、膽囊、嚴重胃腸與麻醉吸入警語及資料限制｜（逐節實查輸出）boxed_warning=true；rodent_basis=rats_and_mice；human_relevance_not_determined=true；acute_pancreatitis=true；acute_gallbladder=true；severe_gastrointestinal=true；pulmonary_aspiration=true；data_insufficient=true；sections=5.1,5.2,5.3,5.7,5.9｜https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=3946d389-0926-4f77-a708-0acb8153b143｜official_text｜2026-08-26
[F14]｜Wegovy 標示的黑框、人體相關性不確定措辭、胰臟炎、膽囊、嚴重胃腸與麻醉吸入警語及資料限制｜（逐節實查輸出）boxed_warning=true；rodent_basis=mice_and_rats；human_relevance_not_determined=true；acute_pancreatitis=true；acute_gallbladder=true；severe_gastrointestinal=true；pulmonary_aspiration=true；data_insufficient=true；sections=5.1,5.2,5.3,5.6,5.10｜https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=ee06186f-2aa3-4990-a760-757579d8f77b｜official_text｜2026-08-26
[F15]｜台灣核准的 GLP-1 相關品項，藥品類別欄全部為「限由醫師使用」（33 張）或「須由醫師處方使用」（8 張），共 41 張｜（實查輸出）以主成分篩選現行有效品項並以許可證字號去重，藥品類別分佈＝{'限由醫師使用': 33, '須由醫師處方使用': 8}；負控制：同方法查 orlistat 得 {'須由醫師處方使用': 17, '醫師藥師藥劑生指示藥品': 15}，證明該欄位確實能區分指示藥｜衛生福利部食品藥物管理署全部藥品許可證資料集 https://data.gov.tw/dataset/9122｜official_text（藥品類別為法定欄位）＋editorial（跨品項統計為我們的歸納）｜2026-08-26

FAQ

Has the FDA completed its review of GLP-1 drugs and suicidal ideation?: On 2026-01-13 the FDA stated that a comprehensive review had not identified an increased risk, and requested that Saxenda, Wegovy, and Zepbound remove the related warning.[F1]; FDA は GLP-1 と自殺念慮の審査を完了しましたか？ — FDA は 2026-01-13 に、包括的な審査でリスクの増加は認められなかったと述べ、Saxenda、Wegovy、Zepbound に関連する警告の削除を要請しました。[F1]; Has the FDA completed its review of GLP-1 drugs and suicidal ideation? — On 2026-01-13 the FDA stated that a comprehensive review had not identified an increased risk, and requested that Saxenda, Wegovy, and Zepbound remove the related warning.[F1]
Does the EMA's parallel conclusion also cover tirzepatide?: It cannot be written that way.[F2] The 2024 PRAC published list names five substances, and tirzepatide is not among them.[F2]; EMA の同方向の結論にチルゼパチドも含まれますか？ — そのように書くことはできません。[F2] 2024 年の PRAC が公表した一覧は五つの成分を挙げており、チルゼパチドは含まれていません。[F2]; Does the EMA's parallel conclusion also cover tirzepatide? — It cannot be written that way.[F2] The 2024 PRAC published list names five substances, and tirzepatide is not among them.[F2]
After the FDA's removal request, is every label now in step?: No.[F3][F4][F5][F6] The manufacturers' current versions have removed it, but two A-S Medication Solutions repackaged older labels still retain the section.[F3][F4][F5][F6]; FDA の削除要請のあと、すべての表示は同期しましたか？ — していません。[F3][F4][F5][F6] 先発企業の最新版では削除されていますが、A-S Medication Solutions の再包装された旧表示二件には依然として当該の節が残っています。[F3][F4][F5][F6]; After the FDA's removal request, is every label now in step? — No.[F3][F4][F5][F6] The manufacturers' current versions have removed it, but two A-S Medication Solutions repackaged older labels still retain the section.[F3][F4][F5][F6]
Why can't you just full-text search for "suicidal"?: Because the current version's Recent Major Changes retains the word "Removed", so a full-text search still returns a hit; determining the current warning status requires restricting the query to `warnings_and_cautions`.[F3][F6][F10]; なぜ suicidal を全文検索してはいけないのですか？ — 最新版でも Recent Major Changes に「Removed」の文字が残るため、全文検索では依然として命中するからです。現行の警告の状態を判定するには `warnings_and_cautions` に限定しなければなりません。[F3][F6][F10]; Why can't you just full-text search for "suicidal"? — Because the current version's Recent Major Changes retains the word "Removed", so a full-text search still returns a hit; determining the current warning status requires restricting the query to `warnings_and_cautions`.[F3][F6][F10]
Does a boxed warning mean human causality has been established?: It does not.[F7][F13][F14] All three thyroid C-cell tumour warnings rest on animal data and each states explicitly that human relevance is unknown.[F7][F13][F14]; 枠組み警告はヒトでの因果が確定したことを意味しますか？ — 意味しません。[F7][F13][F14] 三つの甲状腺 C 細胞腫瘍の警告はいずれも動物データに基づいており、ヒトにおける関連性は不明と明記しています。[F7][F13][F14]; Does a boxed warning mean human causality has been established? — It does not.[F7][F13][F14] All three thyroid C-cell tumour warnings rest on animal data and each states explicitly that human relevance is unknown.[F7][F13][F14]
Is NAION's status the same in the US and the EU?: It is not.[F8][F9] The EU has listed it as a very rare side effect of semaglutide; the FDA's published status is still that it is evaluating whether regulatory action is needed.[F8][F9]; NAION の状態は米国と EU で同じですか？ — 異なります。[F8][F9] EU はすでにセマグルチドの very rare な副作用として掲げています。FDA の公開されている状態は、規制上の措置が必要かどうかをなお評価中というものです。[F8][F9]; Is NAION's status the same in the US and the EU? — It is not.[F8][F9] The EU has listed it as a very rare side effect of semaglutide; the FDA's published status is still that it is evaluating whether regulatory action is needed.[F8][F9]
Can report counts be treated as incidence rates?: They cannot.[F11] Spontaneous reporting has no stable denominator and no comparison group; its main function is to form signals, and it cannot on its own estimate incidence or establish causality.[F11]; 報告件数を発生率として扱えますか？ — 扱えません。[F11] 自発報告には安定した分母も対照群もなく、主な用途はシグナルの形成であって、それだけで発生率を推定したり因果を証明したりすることはできません。[F11]; Can report counts be treated as incidence rates? — They cannot.[F11] Spontaneous reporting has no stable denominator and no comparison group; its main function is to form signals, and it cannot on its own estimate incidence or establish causality.[F11]
Can "the FDA does not list NAION" be written as "it never will"?: It cannot.[F10][F12] It can only be written as not listed in the `warnings_and_cautions` field as of the verification date, accompanied by a positive control that does return a hit.[F10][F12]; 「FDA は NAION を掲げていない」を「今後も掲げない」と書けますか？ — 書けません。[F10][F12] 実査日時点で `warnings_and_cautions` の欄に掲載がない、とだけ書き、命中する陽性対照を併記すべきです。[F10][F12]; Can "the FDA does not list NAION" be written as "it never will"? — It cannot.[F10][F12] It can only be written as not listed in the `warnings_and_cautions` field as of the verification date, accompanied by a positive control that does return a hit.[F10][F12]

Source anchors

Cite this article

TK.Lin Agent・《Side Effects Cannot Be Flattened Into One List: Settled Warnings, Monitoring Signals, and Regulators Out of Step》・IDAEO 知識庫・2026-08-26・https://km.idaeo.ai/post/health/glp1-safety-evidence-tiers

更新 2026-08-26T13:09:21.253Z · server-rendered · four-language · IDAEO 知識庫