跳到内容
现在

健康 · GLP-1眼部警示

FDA要求GLP-1药物加入眼部风险警示:数据意味着什么

全药物类别警示针对的是罕见且可能造成永久性后果的视力丧失。证据值得重视,但不支持一概建议患者放弃治疗。

文字 阅读需 6 分钟 · 24 个来源

原语言视频

译自英文原文。

一项于2024年7月开展的司美格鲁肽研究提出了一个安全性问题,如今,这已成为整个药物类别的警示要求。2026年11月9日,美国食品药品监督管理局(FDA)要求所有GLP-1受体激动剂的生产商——包括Ozempic、Wegovy、Mounjaro和Zepbound——添加有关一种罕见眼部疾病潜在风险增加的信息,该疾病可能造成永久性视力损伤。⁠4

这种疾病是非动脉炎性前部缺血性视神经病变(NAION)。FDA的决定影响重大,但范围比宣称这些药物会导致失明要窄:该决定要求在处方标签的警告和注意事项部分加入相关信息,而不是召回药物或新增黑框警告。该机构尚未规定期限,因此这项公告并不意味着所有标签都已完成更改。⁠4

对患者而言,最明确的指示也最紧迫:突然出现无痛性视力丧失,应立即联系医疗专业人员。盲点、管状视野、色彩失真或中央视力模糊也应立即联系医护人员。不要等到常规处方复诊时再处理。该警示并未要求所有服用这些药物的人停药。⁠4

病例数很少,后果却很严重

当视神经与眼球交界处的前部血流减少时,就会发生NAION。其根本原因尚未完全明了。患者可能在醒来时首先发现一只眼睛视力丧失;损伤可能是永久性的,目前也没有已知的治疗方法。这种疾病罕见但可能造成不可逆伤害,因此无论是危言耸听还是轻描淡写,都无济于事。⁠4⁠1

由于单项研究中的病例很少或没有,FDA汇总了126项随机对照试验。这些试验共纳入134,253名参与者:78,217人接受GLP-1药物,56,036人接受安慰剂或其他降糖药物。以下发生率指随访期间每10万人年可能发生的NAION事件数,而非视力丧失患者所占比例。⁠4

6.4每10万人年可能发生的事件数:GLP-1治疗 4
3.8每10万人年可能发生的事件数:对照治疗 4
15两组试验中的可能事件总数 4

一个人年指一人接受一年的随访。GLP-1组贡献了156,947人年,发生10例可能事件;对照组贡献了131,932人年,发生5例。值得注意的是,所有5例对照组事件都发生在安慰剂使用者中。不过,报告中的3.8这一发生率采用的是整个对照组的随访数据,包括使用其他糖尿病药物的人,而不仅仅是安慰剂使用者。⁠4

未经调整的差值约为GLP-1组每10万人年多出2.6例可能事件。这是描述性相减,并不表示这些病例已被证明由治疗导致。总共只有15例可能事件,这些数据无法给出精确的个人风险,也不能证明每种成分、剂量和用法都具有相同风险。⁠4

作为背景,FDA引用的估计显示,50岁以上成年人每年每10万人中有2.3至10.2例NAION病例。这个范围并非与这些试验匹配的对照组。将它与试验发生率并列,然后断言这些药物安全或危险,是让这项比较承担它无法完成的任务。⁠4

为什么证据仍未定论

据NBC News对FDA审查情况的报道,早期司美格鲁肽研究报告称,糖尿病患者中相关性约为四倍。这一历史发现促使调查展开,但并非目前对整个药物类别风险的估计。FDA指出了若干局限,并将评估范围扩大至临床试验、动物研究、观察性研究,以及药物上市后收集的报告。⁠1⁠4

FDA的另一项Sentinel分析说明,为什么一个引人注目的相对风险数字不足以解决问题。该分析追踪了2019年至2024年间开始使用GLP-1药物或两类其他糖尿病药物——SGLT-2抑制剂和DPP-4抑制剂——的2型糖尿病成年人。GLP-1药物使用者的NAION发生率为每10万人年30至32例,SGLT-2抑制剂使用者为31例,DPP-4抑制剂使用者为26例。这些结果来自不同的研究人群和研究设计,不能替代随机试验中的发生率。⁠4

调整患者群体之间的差异后,与SGLT-2抑制剂的比较得出的风险比为1.19,95%的置信区间为0.91至1.54。该区间包含无差异的可能,因此这一增幅在统计学上并不显著。与DPP-4抑制剂相比,风险比为1.43,区间为1.02至2.01,显示相关性在统计学上升高。只报道第二项结果会掩盖证据中的重要部分。这两项比较都不能证明因果关系。⁠4

FDA还审查了其他严谨的观察性研究,这些研究发现,相较于上述糖尿病治疗药物,相关性有所升高。但个案报告仍难以解读:糖尿病、高血压、吸烟、睡眠呼吸暂停、贫血和肾病本身可能都会增加NAION风险,其中几项在这些药物的处方对象中很常见。所有目前上市的GLP-1产品都曾收到相关病例报告;动物数据审查未发现能够解释药物可能如何增加风险的机制。⁠4

综合来看,证据显示存在值得警示的潜在风险,但并未证明每一例报告病例的成因。全药物类别标签警示要求,是针对这一担忧采取的监管措施,并不证明该类别所有药物的风险完全相同。

哪些药物将受到警示影响

警示范围远不止减重新闻中最熟悉的四个品牌。诺和诺德的司美格鲁肽产品Ozempic、Wegovy和Rybelsus均在其中。礼来公司的Mounjaro和Zepbound也包括在内,它们含有替尔泊肽。替尔泊肽虽然同时作用于GLP-1受体和另一种激素受体GIP,但也被纳入此次行动。2026年获批的奥福格列隆药物Foundayo也在名单之列。⁠4

较早的药物和胰岛素复方制剂也包括在内:含利拉鲁肽的Victoza和Saxenda;含度拉糖肽的Trulicity;将利拉鲁肽与德谷胰岛素合用的Xultophy;将利司那肽与甘精胰岛素合用的Soliqua;以及已上市的仿制药艾塞那肽。FDA表格还列出了历史上的艾塞那肽品牌Byetta,该药于2005年获批。根据2025年8月的一则公告,该药停止销售后,申请人自愿撤回了申请;撤回原因并非安全性或有效性问题。不要将该撤回与此次警示混为一谈。⁠4

这项决定并不是简单地要求停药

这些药物可降低餐后血糖,并作用于控制食欲和进食的脑通路。除用于治疗2型糖尿病和长期体重管理外,某些特定产品还获批用于带来心血管事件、肾病、阻塞性睡眠呼吸暂停或伴有中度至晚期瘢痕形成的代谢性肝病方面的益处。具体用途取决于产品和患者;这些并非所有GLP-1药物都共有的、可以互换的益处。⁠4

既往发生过NAION会增加复发风险,包括另一只眼。曾患NAION的患者应与临床医生讨论是否继续或停止治疗。怀疑发生NAION时,FDA建议临床医生安排及时的眼科评估,并考虑停用该药,除非其益处明确超过风险。⁠4

这意味着有两个时间尺度不同的决策。治疗决策需要根据个人情况权衡利弊。视力突然变化则需要立即采取行动。因果关系尚不确定确实如此,但这并不是延迟评估症状的理由。

来源17

在此处引用的来源中,U.S. Food and Drug Administration最先发布,比本报道播出早32小时。 随后还有11家媒体发布。

5家媒体未注明日期:American Academy of OphthalmologyInvesting.com 香港Medscapeoup.comOvid

按媒体查看所有来源

来自 17 家媒体的 24 个来源· 引用最多的优先

播出原貌
  1. NAION causes sudden, painless vision loss that can be permanent. The FDA now requires warnings about a potential increased risk across all GLP-1 medicines, including widely used diabetes and weight-loss treatments. How large is that risk, and what should patients do about it?
  2. The FDA announced the requirement on November 9, 2026. Manufacturers must add the eye-risk information to the Warnings and Precautions section of prescribing labels. The agency hasn't given a deadline for those changes.
  3. The warning reaches beyond the four familiar names Ozempic, Wegovy, Mounjaro and Zepbound. It covers the full class, including older diabetes medicines, tablets and products that combine a GLP-1 medicine with insulin.
  4. Novo Nordisk's Ozempic, Wegovy and Rybelsus contain semaglutide. Eli Lilly's Mounjaro and Zepbound contain tirzepatide, which acts on both the GLP-1 receptor and another hormone receptor called GIP. Both types are included in the action.
  5. Foundayo, the orforglipron medicine approved in 2026, is also covered. So are Victoza, Saxenda and Trulicity, the insulin combinations Xultophy and Soliqua, and marketed generic exenatide. The requirement spans medicines introduced over more than two decades.
  6. Byetta introduced exenatide in 2005. Its application was voluntarily withdrawn after the brand stopped being marketed, under an August 2025 notice. Safety and effectiveness weren't the reasons for that withdrawal. Generic exenatide remains on the market.
  7. NAION develops when blood flow falls at the front of the optic nerve, where it meets the eye. Its full name is non-arteritic anterior ischemic optic neuropathy. Vision loss can be permanent, and there's no known treatment for the condition.
  8. A person may first notice the change in one eye on waking. Potential risk factors include diabetes, high blood pressure and smoking, along with sleep apnea, anemia and kidney disease. Several of those conditions are already common among people taking these medicines.
  9. Sudden, painless vision loss needs immediate contact with a health care professional. The same advice applies to blind spots or tunnel vision, distorted colors, or blurry central vision. Don't wait for the next routine prescription appointment.
  10. Cases have been reported across all currently marketed GLP-1 products. But those reports are difficult to interpret because patients often have other risks for NAION. The FDA's review of animal data didn't identify how the medicines might trigger the condition.
  11. A July 2024 study found roughly four times the NAION risk among diabetes patients prescribed semaglutide. That early association had several limitations. It prompted a broader FDA assessment of clinical trials, animal research and evidence collected after medicines reached patients.
  12. The pooled trial analysis included 126 randomized trials and 134,253 patients. Of those, 78,217 received GLP-1 medicines and 56,036 received a comparison treatment. Researchers combined the trials because individual studies had few or no cases.
  13. Probable NAION occurred at 6.4 events per 100,000 person-years with GLP-1 treatment, compared with 3.8 in the comparison group. There were just 15 probable events altogether, ten among GLP-1 recipients and five among comparison patients.
  14. A person-year means one person's follow-up for one year. Together, the groups contributed 288,879 person-years. Comparison patients received either placebo or another glucose-lowering medicine. All five comparison events occurred in placebo recipients, but the 3.8 rate uses follow-up from the entire comparison group.
  15. The crude difference works out to about 2.6 more events per 100,000 person-years with GLP-1 treatment. With so few events, that difference can't tell an individual patient their personal risk or show how many cases the medicines caused.
  16. For background, NAION occurs at an estimated 2.3 to 10.2 cases per 100,000 people annually among adults older than 50. Those people aren't a matched comparison group for these trials. A separate real-world analysis also produced different rates.
  17. Adults starting GLP-1 treatment had 30 to 32 events per 100,000 person-years. Rates were 31 with SGLT-2 inhibitors and 26 with DPP-4 inhibitors. This analysis followed adults with type 2 diabetes starting medicines between 2019 and 2024.
  18. SGLT-2 and DPP-4 inhibitors are other types of diabetes medicine. The Sentinel analysis used a different population and study design from the randomized trials. Its rates belong to that separate analysis, and its adjusted comparisons account for differences between patient groups.
  19. The adjusted risk comparison was 1.19 against SGLT-2 inhibitors, with uncertainty spanning no difference. Against DPP-4 inhibitors, it was 1.43, a statistically elevated association. These two comparisons gave different results, and neither proves the medicine caused NAION.
  20. Other rigorous observational studies also found elevated associations against those diabetes treatments. Taken together, the findings led the FDA to require a warning about a potential increased risk. The causal link, the mechanism and the precise risk for particular medicines and patients remain uncertain.
  21. GLP-1 medicines help lower blood sugar after meals and act on brain pathways controlling appetite and food intake. They're used for type 2 diabetes and long-term weight management. For some patients, their approved benefits reach well beyond those uses.
  22. Specific products can reduce major cardiovascular events in certain adults with heart disease, or worsening kidney disease in diabetes. Some also treat moderate to severe obstructive sleep apnea in adults with obesity, or a metabolic liver disease with moderate-to-advanced scarring. Those uses depend on the particular product and patient.
  23. A previous NAION event increases the risk of another, including in the other eye. Patients should discuss continuing or stopping treatment with their clinician. When NAION is suspected, clinicians should arrange prompt evaluation by an ophthalmologist and weigh stopping the medicine against its benefits.

译文更新于

近期播出