这篇专栏描述了「长寿悖论」：有些百岁人瑞抽烟、喝烈酒、每晚喝啤酒、每天吃冰淇淋，甚至每天喝 3 杯 Dr. Pepper，却仍活到 100+ 岁。它提到一项近期研究因主张长寿大约 50% 取决于基因、50% 取决于环境而引发关注，并认为基因占比比许多较早的估计更高。重点不是寿命已被固定，而是对多数人而言，饮食、运动与睡眠仍能实质提高活得更久、更健康的机率；同时也可能有极少数人带有罕见的基因组合，能放慢老化并削弱心脏病、阿兹海默症与癌症在不完美习惯下通常带来的风险。

这项新研究（发表于 Science，并由 Weizmann Institute of Science 主导）使用多个资料集与一个模型，试图滤除较不可能由基因造成的死亡原因（意外、传染病及类似因素），这与一项具影响力、基于数百年家谱纪录的 2018 年前后研究形成对比；后者估计遗传率约为 7%，可能因当时传染病致死普遍而被压低。其他研究则把基因贡献放在约 20%，但研究者主张这些数字会不同，因为它们衡量的终点与年代不同；历史上，早逝使平均寿命远低于 50，而如今美国预期寿命接近 80，尽管基因组并未改变。一个关键主题是：在极端高龄时，基因影响会上升；Thomas Perls 区分活到 80 多岁初期、约 100 岁、以及 105 或 110 岁，并估计 80 多岁中段的长寿约 25% 由基因、75% 由环境决定；到 100 岁约 62% 可归于遗传；而活到更高年龄者则更接近 80%。

补充证据包括一项 2011 年、针对 477 名 96 到 109 岁受试者的研究：他们的行为平均而言比对照组更差，约 50% 曾抽烟、约一半肥胖或过重，且少于 50% 进行哪怕是中等程度的运动；然而他们携带与癌症、阿兹海默症、心脏病与糖尿病相关的基因负荷与他人相近，他们看似的优势反而与较慢老化相关的基因一致。提出的机制包括能抑制促进生长荷尔蒙的「抗老」特征，而像 metformin 与 GLP-1 inhibitors 等药物被提及可能作用于相关途径。专栏提醒：最老年龄层中基因重要性上升，并不否定多数人的生活型态益处；而且环境并非完全由个人掌控，因为长寿与社经地位及空气污染等暴露相关；也指出一个历史模式是，虽然上个世纪预期寿命大幅上升，但活到 100 岁的人所占比例变化不大。作为极端案例，Jeanne Calment 据报在 112 岁才开始抽烟，并又活了约 10 年（到 122 岁），凸显其罕见性，也呼应文章的希望：若能结合更健康的环境与对「顶尖长寿者」的洞见，或许能帮助更多人在健康状态下接近 120 岁。

The column describes the longevity paradox: some centenarians smoke, drink hard liquor, have a beer nightly, eat ice cream daily, or even drink 3 glasses of Dr. Pepper, yet reach 100+ years. It highlights a recent study drawing attention for suggesting longevity is roughly 50% genetic and 50% environmental, a larger genetic share than many earlier estimates. The point is not that lifespan is fixed, but that for most people, diet, exercise, and sleep still meaningfully raise the odds of a longer, healthier life, while a small minority may carry rare gene combinations that slow aging and blunt the usual risks of heart disease, Alzheimer’s disease, and cancer even with imperfect habits.

The new research (published in Science and led by the Weizmann Institute of Science) used multiple datasets and a model intended to filter out deaths from causes less likely to be genetic (accidents, infectious disease, and similar factors), contrasting with influential 2018-era work based on centuries of genealogical records that estimated heritability at about 7%, likely depressed because infectious deaths were common. Other studies have put the genetic contribution around 20%, but researchers argue these figures can differ because they measure different endpoints and eras; historically, early death kept average life expectancy well below 50, while today US life expectancy is nearly 80 even though the genome has not changed. A key theme is that genetic influence rises at extreme ages: Thomas Perls distinguishes getting to the early 80s vs about 100 vs 105 or 110, estimating mid-80s longevity as about 25% genetic and 75% environmental, age 100 as 62% heritable, and survival to even older ages as closer to 80%.

Additional evidence includes a 2011 study of 477 people aged 96 to 109 in which behaviors were, on average, worse than controls: about 50% had smoked, roughly half were obese or overweight, and fewer than 50% did even moderate exercise, yet they carried similar loads of genes associated with cancer, Alzheimer’s disease, heart disease, and diabetes; their apparent advantage instead aligned with genes tied to slower aging. Proposed mechanisms include “anti-aging” profiles that suppress growth-promoting hormones, with drugs like metformin and GLP-1 inhibitors noted as potentially acting on related pathways. The column cautions that rising genetic importance at the oldest ages does not negate lifestyle benefits for most people, and that environment is not fully under individual control because longevity tracks socioeconomic status and exposures like air pollution; it also notes a historical pattern that while life expectancy climbed sharply over the last century, the share reaching 100 changed little. An illustrative extreme case is Jeanne Calment, reported to have started smoking at 112 and lived about another decade (to 122), underscoring both rarity and the article’s hope that combining healthier environments with insights from elite agers could help more people approach 120 in good health.