If you smoke, stopping is the single largest thing you can do for your health — larger than any diet or exercise change in this wiki, and it stays the largest at every age. A lifelong smoker dies at roughly three times the rate of a never-smoker and loses more than a decade of life. Quitting recovers most of that loss, and the earlier it happens the more it recovers: stopping before about 40 removes close to 90% of the excess risk. It is never too late — quitting in one’s fifties still removes most of the harm — but earlier is strictly better, and “safe to smoke until 40” is a misreading. For a non-smoker none of this applies; the lever is already pulled.
Smoking roughly triples the death rate and costs more than a decade of life
A current smoker’s all-cause mortality runs about three times that of a never-smoker: hazard ratio HR 3.0 (99% CI 2.7-3.3) in women and 2.8 (2.4-3.1) in men, after adjustment for age, education, alcohol, and body weight (Jha et al., 2013). (A hazard ratio of 3 means the rate of death at any moment is tripled.) That tripled death rate translates into a large loss of life: «Life expectancy was shortened by more than 10 years among the current smokers, as compared with those who had never smoked» (Jha et al., 2013) — roughly 11 years in women and 12 in men.
The gap shows up in plain survival. Reaching age 80 roughly halves: 70% of never-smoking women get there versus 38% of smokers; for men, 61% versus 26%. About 60% of a smoker’s deaths between ages 25 and 79 are attributable to smoking (Jha et al., 2013).
The excess deaths are cancer, vascular, and respiratory
Those excess deaths concentrate in the diseases smoking is known to cause. Lung cancer carries the steepest relative risk (HR ~15-18), followed by respiratory disease (~9), ischemic heart disease (~3.2-3.5), and stroke (up to ~3.2) (Jha et al., 2013). These are relative rates; because the underlying death rates are already high in older smokers, the absolute toll is heaviest there.
Quitting recovers most of the loss, and the earlier the more it recovers
Cessation is not a marginal adjustment — it returns most of the lost life, and the amount depends heavily on when a smoker stops.
| Quit age | Years of life gained vs continuing | Residual all-cause HR vs never-smoker |
|---|---|---|
| 25-34 | ~10 | 1.0 — essentially a never-smoker |
| 35-44 | ~9 | 1.2 |
| 45-54 | ~6 | 1.5 |
| 55-64 | ~4 | 1.7 |
| continue smoking | — | 2.9 |
Stopping before about 40 removes close to 90% of the excess risk: «Cessation before the age of 40 years reduces the risk of death associated with continued smoking by about 90%» (Jha et al., 2013). Even late cessation pays — quitting at 45-54 still removes roughly two-thirds of the excess risk, and there is no age at which stopping is not the largest available gain.
”Safe to smoke until 40” is a misreading
Quitting near 40 is a floor on the benefit, not a licence. A smoker who stops by about 39 still carries a residual ~20% excess risk (HR 1.2): «about one in six of these former smokers who dies before the age of 80 years would not have died» had they never smoked (Jha et al., 2013). So the “90%” is not a clean reset — earlier is strictly better, and the excess that remains after a late-thirties quit is real.
Why the absolute benefit is so large
The relative effect (HR ~3) sits on a high baseline of smoking-caused death, so the absolute gain is correspondingly enormous — unlike diet levers, whose smaller relative effects sit on lower baselines. The authors make the point directly: «Because the absolute risks of continuing to smoke are large, the absolute benefits of cessation will also be large» (Jha et al., 2013). See Baseline Risk and the Relative-Absolute Split.
Smoking also raises dementia risk — and quitting removes it
Beyond death, smoking is one of the modifiable dementia risk factors, and the 2024 Lancet Commission reclassified it from a late-life to a midlife factor. A 37-study meta-analysis found midlife smoking raised dementia risk (RR 1.30, 1.18-1.45), with «no increased risk in former smokers» (Livingston et al., 2024). So the cessation finding replicates on a second patient-important outcome: quitting removes the excess dementia risk as it removes the excess mortality risk. See Dementia Prevention and Modifiable Risk Factors.
Smoking also raises the risk of rheumatoid arthritis — dose-dependently
Smoking is also a dose-dependent risk factor for rheumatoid arthritis (RA), a chronic autoimmune inflammatory disease. A gold-tier dose-response meta-analysis (10 studies) found incident-RA risk rising with cumulative smoking: even light smokers of 1-10 pack-years carried a 26% higher risk than never-smokers (RR 1.26, 95% CI 1.14 to 1.39), climbing to roughly double the never-smoker risk at heavier exposure — RR 1.94 (1.65 to 2.27) at 21-30 pack-years, RR 2.07 (1.15 to 3.73) above 40 (Di Giuseppe et al., 2014).
The rise plateaus above about 20 pack-years — «the relative risk stabilized approximately at the value of 2» (Di Giuseppe et al., 2014) — but there is no safe low level, since the excess is already significant in the lightest smokers. This evidence is observational (no RA-prevention trial is feasible) and held at low confidence, so RA is a modest add-on to smoking’s harms rather than a headline effect. See Autoimmune Disease and Modifiable Risk.
Why this is held with high confidence — despite being observational
No one randomizes smoking, so the evidence is observational. But the read is decisive here for reasons that a trial would not improve on: an HR of ~3 is far too large to be confounded away; adjustment for education, alcohol, and adiposity «did little to alter the hazard ratios» (Jha et al., 2013); and the result is concordant across four independent cohorts (British doctors, the Million Women Study, a 50-year US series, and a Japanese cohort). The analysis also handles the sick-quitter bias — that dying smokers quit and inflate the apparent risk of recent quitters — by reclassifying anyone who quit within five years of death as a current smoker, which biases against the benefit. The large benefit survives that conservative handling.
What the held evidence does not yet cover
The fabric holds one gold mortality source (Jha 2013) plus Livingston 2024 for dementia. The coverage-gap inventory below is the wiki’s own appraisal of what the held sources omit, not a claim from either source. The inventory is honest about what it does not reach:
- E-cigarettes / vaping — the cardiovascular cell is now open
(Electronic Cigarettes and Cardiovascular Risk) but only on surrogates; hard patient-important
CV outcomes remain insufficient evidence, and the decision is stratum-dependent — an e-cigarette is
not one thing, and the sign flips with who uses it:
- A smoker who switches (vs continued combustible smoking) gets a harm-reduction surrogate benefit — blood pressure fell: SBP -7.00 mmHg (95% CI -9.63 to -4.37), DBP -3.65 mmHg (-5.71 to -1.59) (Skotsimara et al., 2019). But hard CV outcomes stay insufficient evidence: the one observational MI estimate (OR 1.79) sits below combustible cigarettes (OR 2.72) yet is confounded, with reverse causation flagged (Skotsimara et al., 2019) — so the surrogate improvement is held, but a hard-outcome benefit is not demonstrated.
- A never-smoker who starts (vs no exposure) takes on pure added harm with no offsetting baseline: acute use raised heart rate +2.27 bpm, SBP +2.02 mmHg, DBP +2.01 mmHg (Skotsimara et al., 2019), with no cessation benefit to offset it. For a smoker who cannot quit, switching is a plausible harm-reduction step on surrogates whose hard-outcome case is unproven; for a never-smoker, starting is unambiguous added risk.
- Secondhand smoke — no dose-response on hard outcomes is held.
- Specific cancers beyond mortality — no dedicated effect-size meta-analyses (e.g. smoking and a named cancer) beyond Jha’s mortality figures.
- How to quit — the comparative efficacy of cessation methods (nicotine replacement, pharmacotherapy, behavioural support) is out of scope by design: those are prescriber and treatment-selection acts. This deliverable appraises the value of quitting, not the method.
Bottom line — what to do
- If you smoke, quitting is the top priority — ahead of every diet, exercise, or sleep change. It is the #1 modifiable lever whenever it is present -> Layer 1 - Ranking Interventions for a Stratum.
- Quit as early as possible. Before ~40 removes close to 90% of the excess risk; a residual ~20% remains even then, so sooner is genuinely better.
- It is never too late. Quitting in one’s forties or fifties still removes most of the excess risk and adds years of life.
- The benefit shows up on more than one outcome — mortality, life expectancy, and dementia all move the same way with cessation.
- For a non-smoker, this lever is already pulled — the question moves to the next-largest exposure.
Evidence box
Question What does the evidence show about smoking and patient-important outcomes — how large is the effect on mortality, life expectancy and specific diseases, how much of any excess risk cessation recovers and how that depends on the age at quitting, and what remains uncertain? Evidence included 4 sources — 2 gold, 1 high, 1 moderate Overall certainty High (see Rating Certainty of Evidence) Source-selection note Core mortality/cessation evidence is gold or high tier; the e-cigarette note rests on one moderate, surrogate-only meta-analysis and is held at low confidence. Last updated 2026-09-03 · Independently reviewed: No · Full edit history