Smoking & Quitting

What the evidence actually shows

Smoking is the cleanest large effect in the entire longevity literature, which is a grim thing to be able to say. Jha and colleagues, using US national survey data linked to death records, found that current smokers lost roughly a decade of life expectancy compared with never-smokers, and that the all-cause mortality rate ratio for smokers versus never-smokers sat around 2.5 to 3 across middle age. The British Doctors Study — Doll and Peto following physicians for fifty years — reached the same neighbourhood from an entirely different cohort in a different country in a different era. When two independent designs land on the same number, you stop arguing about confounding.

The reason the smoking signal is trustworthy where other lifestyle signals are muddy is dose and specificity. The effect is enormous, it is dose-dependent in pack-years, it shows up in specific causes with known biological pathways (lung cancer, COPD, vascular disease), and it reverses on cessation. That combination is about as close to causal proof as epidemiology gets without a randomised trial nobody would ever run.

How big the effect really is, in years

Roughly ten years of life expectancy for a lifelong smoker versus a never-smoker. That is not hedged much, because the data does not require it. What does deserve hedging is the recovery figure, and here the news is genuinely good: quitting is not a consolation prize.

Jha's data show that quitting before about age 40 avoids the great majority — on the order of 90 percent — of the excess mortality that continued smoking would have caused. Quit at 50 and you still recover roughly half. Quit at 60 and you recover something, but the accountant in me will tell you the compounding has already happened. The single most valuable sentence a smoker can hear is that the clock genuinely runs backwards on cessation, and that it runs backwards fastest for the young.

So the honest year-figures are: about −10 for continued smoking, but the recoverable fraction is a steep function of the age you stop. This is one of the rare places where the intervention's upside is nearly as large as the exposure's downside.

The catch

The main caveat is not about whether smoking kills — it is about the shape of the recovery curve and about what "former smoker" means. A former smoker's risk stays elevated above never-smoker for years, decaying toward baseline over roughly a decade or two depending on how long and how heavily they smoked. So "I quit" is not a single state; it is a trajectory, and where you sit on that trajectory depends on both pack-years accumulated and years since quitting.

The other honest note is that residual risk never fully returns to never-smoker for heavy long-term smokers, particularly for lung cancer. The curve gets close but does not perfectly touch. We would rather tell you that than pretend the slate wipes clean, because false reassurance is its own kind of harm.

How this feeds your actuarial age

In the calculator the smoking modifier is the heaviest single behavioural weight, and deliberately so. We distinguish current, former, and never, and for former smokers the adjustment decays with years since quitting rather than snapping to zero the day you stub out the last one — because that is what the mortality data actually does.

If you are a current smoker, this modifier will move your actuarial age more than anything else on the site, and no amount of zone-2 cardio, cold plunging, or supplement stacking compensates for it. That is not a moral judgement; it is just where the hazard ratios are. If you are considering one change, this is the one that pays. As always, on the practical how — nicotine replacement, medication, whatever works for you — consult your clinician.