🧬 Aging Wiki

smoking

Properties1
aliasessmoking, tobacco smoking, cigarette smoking, tobacco use, cigarettes

14 min read

Tobacco Smoking

Tobacco smoking is the leading single modifiable cause of premature death globally, responsible for approximately 6–8 million deaths per year. Lifelong smokers lose roughly one decade of life expectancy compared to never-smokers 1; the 50-year British Doctors Study established that men born in the 1920s who smoked persistently had approximately threefold higher age-specific mortality than never-smokers, dying on average about 10 years younger 2. At the cellular level, smoking accelerates virtually every established hallmark of aging — driving genomic damage via DNA adducts, epigenetic-clock acceleration (the strongest non-cancer lifestyle modulator of clock-based biological age), NF-κB–mediated chronic inflammation, cellular senescence, mitochondrial dysfunction, and telomere attrition. Cessation yields large, time-dependent risk reductions.

Epidemiology and mortality burden

Lifetime risk and life-expectancy loss

The 50-year follow-up of the British Doctors cohort (Doll et al. 2004; n=34,439 men; initiated 1951) found that mortality hazard in men born 1900–1909 was approximately twofold higher in smokers vs never-smokers (42% vs 24% probability of dying in middle age, ages 35–69), rising to threefold for men born in the 1920s (43% vs 15%) as cumulative exposure increased 2. Men born around 1920 who smoked throughout adult life died on average about 10 years younger than lifelong non-smokers. Cessation at age 50 halved the excess risk; cessation at age 30 avoided almost all of it over subsequent decades 2.

The US Cancer Prevention Study analysis (Jha et al. 2013; NEJM; >100,000 person-years of prospective follow-up) confirmed that current smokers have approximately 3× the all-cause mortality hazard of never-smokers and lose at least 10 years of life expectancy on average 1. Cessation before age 40 reduced the risk of death associated with continued smoking by about 90% 1.

Dose-response

Risk scales with pack-years: risk is monotonically increasing with cumulative dose without a clearly established safe threshold for all-cause mortality. Occasional or light smoking (~1–4 cigarettes/day) still carries a relative risk of all-cause mortality approximately 1.5-fold above never-smokers in large-cohort analyses, underscoring the absence of a safe threshold. gap/dose-response-unclear (safe threshold has not been formally established via MR).

Mendelian randomization evidence

Smoking is among the best-instrumented behavioral exposures in MR. Key instruments:

  • CHRNA5-CHRNA3-CHRNB4 locus (rs16969968 / rs1051730): the most-replicated smoking-intensity instrument in the nicotinic acetylcholine receptor gene cluster on 15q25.1. Used across multiple MR studies to instrument cigarette consumption among smokers. (Distinct from CHRNA7/α7nAChR on 15q13.3 — a different nicotinic-receptor gene at a different locus, mediating the anti-inflammatory cholinergic reflex rather than nicotine-dependence intensity.)
  • GSCAN consortium (Liu et al. 2019, Nature Genetics; n up to 1.2 million): identified 566 genome-wide significant variants across 406 loci for smoking initiation, cessation, heaviness, and alcohol use; these instruments are widely used for smoking-exposure MR in subsequent work 3.

Selected MR-supported causal inferences:

OutcomeInstrumentResultReference
Telomere attrition377 smoking variants (UKB)Smoking genetically associated with shorter leucocyte telomere length (β = −0.041, 95% CI −0.054 to −0.028) and reduced appendicular lean mass; causal links to sarcopenia phenotypes4
Physical function declinePolygenic smoking scoreCausal association with impaired body mobility (β = 5.553; 95% CI 1.029–10.077) and reduced ADL index (β = 1.908; 0.196–3.619) in 29,139 UK older adults5
Depression, schizophreniaGSCAN smoking initiation + UKB lifetime smokingBidirectional MR supports causal effect of smoking on both conditions6

Smoking-MR instruments have been validated against positive-control outcomes including lung cancer. Two-sample MR using GSCAN instruments consistently shows larger effect-size estimates than observational analyses for multiple aging-related endpoints, consistent with confounding attenuating the observational signal.

Aging mechanisms and target hallmarks

1. Genomic instability — DNA adducts and strand breaks

Tobacco smoke contains >70 established carcinogens, including polycyclic aromatic hydrocarbons (PAHs such as benzo[a]pyrene), nitrosamines (NNK, NNN), and benzene. These form bulky DNA adducts that, if unrepaired, cause substitution and insertion/deletion mutations. Smokers’ lung tissue and blood carry substantially higher adduct burdens than never-smokers, and mutational signatures attributable to tobacco (COSMIC Signatures 4, 29, 92) are detectable by whole-genome sequencing. gap/needs-replication for quantitative adduct-per-pack-year estimates in non-respiratory tissues.

At the population level, smoking is the dominant environmental mutagen driving somatic mutation accumulation — implicated in causation of at least 14 distinct cancer types beyond lung cancer.

2. Epigenetic alterations — epigenetic-clock acceleration and AHRR methylation

Smoking produces the strongest non-disease lifestyle signal in epigenetic-clock analyses. Key loci:

  • cg05575921 (AHRR): hypomethylation at this CpG in the aryl hydrocarbon receptor repressor gene is the most replicated smoking-associated methylation change, present in blood, airway epithelium, and other tissues. Partial reversal occurs on cessation.
  • GrimAge acceleration: the GrimAge second-generation epigenetic clock includes DNAmPackYears as one of its seven component plasma-protein surrogates. Current smokers show ~6.36-year GrimAge acceleration (95% CI 5.14–7.58) relative to never-smokers; former smokers retain 2.34-year residual GrimAge acceleration (95% CI 1.66–3.02) relative to never-smokers 7. Horvath pan-tissue EAA (a distinct clock) is elevated by 1.69 years (95% CI 0.72–2.67) in former smokers vs never-smokers but not significantly elevated in current smokers in this cohort 7. GrimAge (incorporating smoking signal) is among the strongest DNAm-based predictors of all-cause mortality 8.
  • Airway and lung tissue epigenetic aging: smoking accelerated epigenetic age by ~4.9 years in airway cells and ~4.3 years in lung tissue; airway reversal was observed after cessation, lung tissue showed more persistent effects 9.
  • Early-life effects: gestational and childhood tobacco smoke exposure accelerate epigenetic aging in infants and children 10, indicating trans-generational and developmental-programming mechanisms.

3. Cellular senescence

Tobacco smoke constituents — particularly acrolein, hydrogen cyanide, and reactive oxygen species from tar combustion — induce premature cellular senescence in lung epithelial cells, fibroblasts, and endothelial cells. Senescent cells accumulate in smokers’ airways and are a key driver of the SASP (senescence-associated secretory phenotype) that propagates chronic airway inflammation 11.

Smoke-induced autophagy impairment is a proximal mechanism: reactive species damage selective autophagy receptors (p62/SQSTM1), impairing clearance of damaged organelles and forcing cells into senescence 11. In lung cancer, the resulting SASP creates a pro-tumorigenic microenvironment, explaining the epidemiologically observed synergy between smoking-induced senescence and carcinogenesis 12.

4. Chronic inflammation — NF-κB activation

Tobacco smoke activates nf-kb via multiple arms: (i) direct oxidant activation of IκB kinase, (ii) toll-like receptor stimulation by lipopolysaccharide contaminants in tobacco, (iii) NLRP3 inflammasome activation by particulate matter. The resulting cytokine cascade — elevated IL-6, IL-8, TNF-α, CRP — is measurable as a population-level inflammaging signal in smokers compared with never-smokers. The SASP from senescent airway cells amplifies this signal. On cessation, systemic inflammatory markers decline within months, though some degree of elevation may persist for years. gap/long-term-unknown (degree of residual NF-κB activation after >10 years cessation not quantified in humans).

Note a countervailing nuance: nicotine non-selectively also activates α7nAChR, engaging the cholinergic-anti-inflammatory-pathway, which can produce acute anti-inflammatory/immunomodulatory effects in experimental systems. This does not offset smoking’s harms — chronic tobacco exposure is strongly net pro-inflammatory (and receptor desensitization blunts any cholinergic brake) — but it is why isolated α7 agonism is pursued as an anti-inflammatory strategy independent of smoking.

5. Mitochondrial dysfunction

Tobacco smoke mitochondrial toxins — particularly carbon monoxide (competitive hemoglobin binding, reducing O₂ delivery) and hydrogen cyanide (complex IV inhibition) — acutely impair oxidative phosphorylation. Chronically, reactive oxygen species from smoke impair mitochondrial membrane integrity, drive mtDNA mutation accumulation, and reduce respiratory chain efficiency. Mitochondrial biogenesis signaling (PGC-1α / TFAM axis) is suppressed. These effects are partially reversed on cessation as mitochondrial biogenesis can recover. gap/needs-replication for quantitative mtDNA copy number or mitochondrial membrane potential data stratified by pack-years in human studies.

6. Telomere attrition

MR analysis (Park et al. 2023; n=337,138; 377 genetic variants) provides causal evidence that smoking shortens leucocyte telomere length (β = −0.041 SD; 95% CI −0.054 to −0.028) 4. Observational meta-analyses consistently report shorter telomeres in smokers vs never-smokers, with an approximately 4.6 bp per pack-year attrition estimate, though this meta-analytic figure requires MR verification for causality. The telomere-attrition signal partially underpins the sarcopenia and reduced muscle mass outcomes in smoking MR work 4.

Downstream phenotypes

PhenotypePopulation RR (smokers vs never)Notes
cancer (lung)RR ~20–25×Largest attributable fraction of any behavioral exposure for lung cancer
cancer (14+ other sites)RR 1.5–5× by siteBladder, kidney, esophagus, pancreas, cervix, stomach, mouth, larynx among most attributable
atherosclerosis / CVDRR ~2–3× for CHDEndothelial dysfunction, oxidative LDL modification, platelet activation
copdRR ~10–15× lifetimeDominant modifiable cause; smoke-driven airway senescence is the proximal pathology
type-2-diabetesRR ~1.3–1.4×Nicotine-mediated insulin resistance + adipose inflammation
chronic-kidney-diseaseRR ~1.3–2.0×Endothelial dysfunction + direct tubular toxicity from cadmium/lead in tobacco
Sarcopenia / muscle lossCausal MR evidence (see above)Telomere attrition + systemic inflammation + reduced physical activity

Evidence quality: cancer (lung), atherosclerosis, copd — strong (large RCT-equivalent prospective evidence; some MR confirmation); type-2-diabetes, chronic-kidney-disease — limited to moderate (large observational, partial MR support).

Reversibility and cessation benefit

Quantitative time-course

Time after quittingRisk reduction (vs continued smoking)
20 minutesHeart rate and blood pressure fall toward normal
24 hoursCarbon monoxide cleared; blood O₂ levels normalize
1 yearExcess coronary heart disease risk approximately halved
5–15 yearsStroke risk returns to that of never-smokers (Jha 2013 data)
10+ yearsLung cancer mortality risk roughly halved vs continued smokers
20+ yearsAll-cause mortality excess vs never-smokers substantially reduced; complete equalization not achieved for lung cancer and COPD

Jha et al. 2013 established that adults who quit at ages 25–34 gained about 10 years of life expectancy (recovering essentially all of the loss); those who quit at ages 35–44 gained about 9 years; those who quit at ages 45–54 gained about 6 years. Cessation before age 40 reduced the risk of death associated with continued smoking by about 90% 1.

Asian pooled-cohort data (Yang et al. 2022; n across 16 cohort studies) found that all-cause and CVD mortality remained elevated among former smokers at 10–14 years post-cessation (HR ~1.25 and 1.20 respectively vs never-smokers), and lung cancer mortality remained approximately 2-fold higher than never-smokers 15–19 years post-cessation — supporting the view that residual risk persists for >2 decades 13.

Epigenetic reversibility

Airway epigenetic aging (assessed by Horvath clock on bronchial cells) shows partial reversal after cessation 9. AHRR (cg05575921) hypomethylation partially restores with cessation duration, and is used as a molecular marker of cessation success. GrimAge acceleration in former smokers (2.34-year excess vs never-smokers 7) persists beyond biological-age normalization observed for AHRR alone, suggesting that some epigenetic damage is more durable. Horvath pan-tissue EAA shows a smaller but still-significant residual in former smokers (1.69 years 7).

Extrapolation and model-organism notes

DimensionStatus
Pathway conserved in humans?yes — all mechanistic pathways (DNA adducts, NF-κB, AHRR methylation) studied in human cells/tissue
Phenotype conserved in models?yes — mouse and rat inhalation exposure models recapitulate airway senescence, inflammatory response, and lung pathology
Causal evidence in humans?yes — large prospective cohorts + MR instruments confirm causality

The primary evidence base for smoking is human-derived (British Doctors, US CPS, UK Biobank, GSCAN). Animal models are used mechanistically but the epidemiological foundation is uniquely human.

Limitations and knowledge gaps

  • No safe threshold established: while dose-response is monotonic, the MR-confirmed lower boundary for where risk converges to never-smoker levels has not been formally established. Observational “light smoking” data suggest no safe threshold, but MR instruments instrument intensity among current smokers, not casual vs never. gap/dose-response-unclear
  • Epigenetic reversibility quantification: the degree and tissue-specificity of epigenetic-clock reversal after sustained cessation (>10 years) requires larger longitudinal MR-embedded studies. GrimAge residual acceleration in former smokers 7 needs replication with longer follow-up. gap/needs-replication
  • Senescence causal chain: while smoke-induced senescence is documented in airway tissue, the causal contribution of local senescent cells (vs systemic SASP) to downstream phenotypes like COPD is not yet resolved. gap/no-mechanism
  • Mitochondrial recovery kinetics: quantitative mtDNA heteroplasmy and respiratory capacity data post-cessation in humans are sparse. gap/long-term-unknown
  • COPD and CKD phenotype pages: copd and chronic-kidney-disease are implicit stubs; these pages should be seeded to complete the downstream-phenotype linkage. stub
  • Sex differences: the epidemiological magnitude of smoking risk differs between males and females (women may have higher per-cigarette lung cancer risk at equivalent doses; not yet reconciled with the sex-differential aging literature). gap/needs-replication

Footnotes

Footnotes

  1. doi:10.1056/NEJMsa1211128 · Jha P, Ramasundarahettige C, Landsman V, Rostron B, Thun M, Anderson RN, McAfee T, Peto R · N Engl J Med 2013;368:341–350 · prospective cohort · n=202,248 (113,752 women + 88,496 men; US National Health Interview Survey 1997–2004, followed to 2006); current smokers HR ~3.0 (women) and ~2.8 (men) vs never-smokers; smokers lose >10 yr life expectancy; quitting at 25–34 gains ~10 yr, at 35–44 gains ~9 yr, at 45–54 gains ~6 yr; cessation before age 40 reduces mortality risk by ~90% · PMID 23343063 · abstract-verified (full-text PDF pending download) ↩ ↩2 ↩3 ↩4

  2. doi:10.1136/bmj.38142.554479.AE · Doll R, Peto R, Boreham J, Sutherland I · BMJ 2004;328:1519 · prospective cohort · n=34,439 male British doctors followed from 1951; 50-year follow-up · men born 1900–09: twofold death-rate ratio (middle age 42% vs 24%); men born 1920s: threefold (43% vs 15%); persistent smokers born ~1920 died ~10 yr younger than non-smokers; cessation at 50 halved the excess mortality; cessation at 30 avoided almost all of it · PMID 15213107 · abstract-verified (full-text PDF pending download) ↩ ↩2 ↩3

  3. doi:10.1038/s41588-018-0307-5 · Liu M, Jiang Y, Wedow R, Li Y, Brazel DM, Chen F, et al. (GSCAN Consortium) · Nat Genet 2019;51(2):237–244 · GWAS meta-analysis · n up to 1.2 million; 566 variants across 406 loci for tobacco and alcohol use combined (smoking initiation 378 variants, cigarettes/day 55, cessation 24, age of initiation 10; alcohol 99); canonical source for MR instruments in subsequent smoking-exposure studies · PMID 30643251 ↩

  4. doi:10.1002/jcsm.13174 · Park S, Kim SG, Lee S, Kim Y, Cho S, Kim K, et al. · J Cachexia Sarcopenia Muscle 2023;14(1):80–92 · Mendelian randomization · n=337,138 UK Biobank participants; 377 genetic smoking variants; causal association with shorter leucocyte telomere length (β=−0.041; 95% CI −0.054 to −0.028) and with sarcopenia phenotypes (reduced muscle mass, slower walking pace) · PMID 36696951 ↩ ↩2 ↩3

  5. doi:10.1136/jech-2021-217572 · Gaggero A · J Epidemiol Community Health 2022;76(6):554–560 · Mendelian randomization · n=29,139 UK older adults (mean age 65.8 yr); polygenic smoking score instruments; causal association with impaired body mobility (β=5.553; 95% CI 1.029–10.077) and reduced ADL index · PMID 35145020 ↩

  6. doi:10.1017/S0033291719002678 · Wootton RE, Richmond RC, Stuijfzand BG, Lawn RB, Sallis HM, et al. · Psychol Med 2020;50(14):2435–2443 · Mendelian randomization · GSCAN + UK Biobank instruments; bidirectional MR supports causal effects of smoking on depression and schizophrenia · PMID 31689377 ↩

  7. doi:10.1038/s41598-022-08160-w · Cardenas A, Ecker S, Fadadu RP, Huen K, Orozco A, McEwen LM, et al. · Sci Rep 2022;12:4277 · observational EWAS cohort · n=489 Costa Rican adults (mean age 79.4 yr; 7.6% current, 35% former, 57.4% never smokers); current smokers: 46 differentially methylated CpGs (AHRR, F2RL3 regions), GrimAge acceleration 6.36 yr (95% CI 5.14–7.58) vs never-smokers; former smokers: GrimAge acceleration 2.34 yr (95% CI 1.66–3.02), Horvath EAA 1.69 yr (95% CI 0.72–2.67), shorter DNAmTL (−0.04 kb; 95% CI −0.08 to −0.01) vs never-smokers · PMID 35277542 · PMC8917214 ↩ ↩2 ↩3 ↩4 ↩5

  8. doi:10.18632/aging.101684 · Lu AT, Quach A, Wilson JG, Reiner AP, Aviv A, Raj K, Hou L, Baccarelli AA, Li Y, Stewart JD, Whitsel EA, Assimes TL, Ferrucci L, Horvath S · Aging (Albany NY) 2019;11(2):303–327 · observational, multi-cohort · GrimAge composite incorporates DNAmPackYears surrogate; GrimAge strongly predicts all-cause mortality and is an epigenetic integrator of cumulative smoking exposure · PMID 30669119 ↩

  9. doi:10.1186/s13148-019-0777-z · Wu X, Huang Q, Javed R, Zhong J, Gao H, Liang H · Clin Epigenetics 2019;11:119 · observational cohort + cessation comparison · smoking accelerated epigenetic age +4.9 years in airway cells and +4.3 years in lung tissue; airway reversal observed post-cessation; lung tissue effects more persistent · PMID 31801625 ↩ ↩2

  10. doi:10.1016/j.envint.2021.106683 · de Prado-Bert P, Ruiz-Arenas C, Vives-Usano M, et al. · Environ Int 2021;155:106683 · observational birth cohort · n=1,173 children; >100 early-life exposures; gestational tobacco smoke exposure and childhood indoor PM associated with accelerated epigenetic aging in children · PMID 34144479 ↩

  11. doi:10.1002/jbt.70065 · Akhtari M, Jalalvand M, Sadr M, Sharifi H · J Biochem Mol Toxicol 2024;38(12):e70065 · review · tobacco smoke induces cellular senescence via oxidative-stress-driven autophagy impairment; SASP amplifies airway inflammation and promotes lung tumorigenesis · PMID 39588771 ↩ ↩2

  12. doi:10.1016/j.arr.2024.102315 · Jha SK, De Rubis G, Devkota SR, Zhang Y, Adhikari R, Jha LA, Bhattacharya K, Mehndiratta S, Gupta G, Singh SK, Panth N, Dua K, Hansbro PM, Paudel KR · Ageing Res Rev 2024;97:102315 · review · cellular senescence in lung cancer; SASP as both tumor-suppressor and pro-oncogenic driver in smoke-exposed lung; therapeutic targeting landscape · PMID 38679394 ↩

  13. doi:10.1093/ije/dyab087 · Yang JJ, Yu D, Shu X-O, Wen W, Rahman S, et al. · Int J Epidemiol 2022;51(1):318–331 · pooled cohort · 16 Asian population cohorts; all-cause and CVD mortality remain elevated in former smokers at 10–14 years post-cessation (HR ~1.25 and 1.20); lung cancer mortality ~2× higher vs never-smokers at 15–19 years post-cessation · PMID 34999862 ↩

Graph View

Backlinks