Acarbose

Acarbose was approved in the 1990s to treat type 2 diabetes. It works in the gut, not the blood: by blocking the enzymes that split starches and sucrose into simple sugars, it flattens the glucose spike after a carbohydrate-heavy meal. That mechanism made it an early candidate for the theory that repeated post-meal glucose spikes contribute to aging.

The strongest data behind that theory comes from mice, not humans. The National Institute on Aging's Interventions Testing Program (ITP), which runs the same compound across three independent labs and multiple genetically diverse mouse strains, found that acarbose extended median male lifespan by 22% at a dose of 1,000 ppm in its original cohort [1]. A follow-up ITP study testing three doses (400, 1,000, and 2,500 ppm) confirmed the effect but found no extra benefit at higher doses, and a clear sex gap: the two higher doses produced 16-17% increases in male median lifespan versus only 4-5% in females [2]. The sex difference wasn't explained by body weight or fat mass, which dropped more in females than males on the drug. Among the compounds the ITP has tested, acarbose remains one of a handful with a reproducible, multi-site, multi-strain lifespan signal [3].

What that mouse result does not show is a human aging benefit. Acarbose has decades of use in type 2 diabetes, where trials measure blood glucose, HbA1c, and cardiovascular events, not lifespan or biological age. A meta-analysis of acarbose in Chinese patients with diabetes or prediabetes found reasonable glycemic efficacy but did not evaluate longevity outcomes [4]. No RCT has tested acarbose against a biological aging clock or all-cause mortality in generally healthy people. The ITP itself only tests genetically heterogeneous mice, and translating a 22% median-lifespan gain in mice to any specific human effect size is not something the data support.

The more immediate problem for anyone considering acarbose off-label is tolerability. Acarbose works by deliberately letting undigested carbohydrate reach the colon, where gut bacteria ferment it. The predictable result is gas, bloating, and diarrhea; these gastrointestinal effects are the most common reason patients discontinue the drug in clinical practice, and they are dose-dependent [5]. A drug that most people won't take at an effective dose because of GI distress is a harder sell as a longevity intervention than a mouse hazard-ratio curve suggests. Separately, small rodent studies report acarbose reduces age-related tissue lesions in heart and kidney of C57BL/6 mice, a mechanistic finding consistent with a metabolic slowing effect, but again mouse tissue, not a human outcome [5].

Bryan Johnson's public Blueprint protocol does not list acarbose among his current or former supplements; he has instead emphasized meal sequencing and a fixed low-glycemic diet to manage post-meal glucose, which is a behavioral analog to the same theory rather than the drug itself.

Critics of extrapolating ITP results to human longevity, including gerontologists who work directly with the program, note that mouse lifespan studies systematically fail to predict human drug effects because mice die of different diseases (mostly cancer) than humans do, and because ITP cohorts are not calorie-restricted or exercised, unlike free-living humans, which changes baseline metabolic stress. The GI-tolerability gap is a separate, more practical critique: even if the biology translates, dosing that most people will actually sustain may be below the effective mouse-equivalent dose.

Acarbose has the most consistent multi-site mouse lifespan data of any repurposed diabetes drug in longevity research, but zero human trials test it for aging, and its main real-world limitation is that most people stop taking it once GI side effects set in.