Yeon Wook Kim and colleagues’ new paper “Electronic cigarette use after smoking
cessation and lung cancer risk” shows that smokers who switch to e-cigarettes have about a 50% higher risk of developing lung cancer and about a doubling of risk of dying from lung cancer than smokers who just stop smoking.
Most important, this paper is direct high-quality evidence against e-cigarette advocate and tobacco company claims that switching from cigarettes to e-cigarettes would reduce lung cancer risk because e-cigarettes expose users to lower levels of carcinogenic combustion products in tobacco smoke, the central claim behind the argument that e-cigarettes are an effective harm reduction strategy.
The study itself is a real tour de force. It based on essentially the entire population of South Korea because it is based on relevant people enrolled in the South Korean National Health Insurance Service (NHIS), South Korea’s single-payer universal health system that provides mandatory health insurance for all citizens. In particular, NHIS covers a biennial health screening program for all employees, including self-employed insured individuals aged ≥20 years, and everyone aged ≥40 years regardless of
employment status. The NHIS collects information on lifestyle behaviors, including smoking and e-cigarette use. It covers nearly all forms of health care and maintains the national records of all health screening results, demographic characteristics, diagnoses, inpatient and outpatient visits, medical procedures and pharmaceutical prescriptions of 97% of the South Korean population.
The analysis is based on 4,524,895 adults who participated in the Korean National Health Screening Program who did not have lung cancer in 2018 (baseline), with prior records from 2012-2014. Participants were classified as current smokers, short-term quitters or long-term quitters, and followed up for 6 years, to December 2023. Over 24,182,543 person-years, 35,887 lung cancers and 12,807 lung cancer deaths occurred. In addition to being a gigantic and universal sample, the longitudinal design is the strongest kind of epidemiological study because it ensures that the exposure (e-cigarettes) occurred before the development if disease (lung cancer), which avoids concerns about temporality present in cross-sectional studies.
The key finding: Compared with complete quitters, daily e-cigarette use after smoking cessation was associated with higher risks of lung cancer incidence (adjusted hazard ratio (aHR) 1.56, 95% confidence interval (CI) 1.24-1.97) and lung cancer death (aHR 2.00, 95% CI 1.28-3.15).
The analysis controlled for age, sex, smoking intensity in pack years, chronic obstructive pulmonary disease, pulmonary fibrosis, other diseases, body mass index (BMI), income, and physical activity. The results of the fully adjusted model were similar to a simpler model that only included age, sex and pack years of smoking. This finding, which is similar to studies of the associations of e-cigarettes with many other diseases, suggests that, despite claims by e-cigarette advocates, residual confounding by other variables is unlikely to explain the observed link between e-cigarette use and lung cancer among former smokers.
Results were also similar using any e-cigarette use in the past month as the independent variable.
Associations were consistent in short-term and long-term quitters and even higher in the high-risk subgroup (people aged 50–80 years with greater than 20 pack years of smoking; incidence: aHR 1.91, 95% CI 1.44-2.53; death: aHR 1.92, 95% CI 1.13-3.24).
The authors also did many additional analyses using different comparisons, which all supported the overall conclusion that e-cigarette use after quitting smoking increased lung cancer risk, further increasing the confidence we can have in the paper.
Of course, an e-cigarette advocate dismissed the study
A news story about the study In the Daily Mail quoted Professor Peter Hajek, “an expert in clinical psychology and lifestyle research, who wasn’t involved in the current study, [who] raised concerns about the interpretation of the results.”
- “In the group where the difference in lung cancer was the largest, non-vapers have been smoke-free decades longer than vapers.
- “In addition, an unknown but possibly substantial proportion of vapers with lung cancer are likely to have picked up vaping as means of stopping smoking after they were diagnosed with lung cancer.
- “This raises a concern about misinterpreting these results as suggesting that vaping causes lung cancer.
- “Such reporting can by itself increase the risk of lung cancer if it discourages smokers from switching to a much less risky product.”
Reading these rather generic criticisms that e-cigarette advocates raise about almost every study leads me to wonder if Hajek actually read the full paper, since Kim and colleagues anticipated all these criticisms and showed that they were unlikely to be real problems.
With regard to changes in smoking status, Kim and colleagues reported:
To account for changes in smoking status and e-cigarette use after baseline, an additional sensitivity analysis was conducted in a subpopulation of participants who underwent additional health screening during 2019–2020 with updated information on smoking and e-cigarette use. Using this exploratory subpopulation (n = 3,465,275), time-varying Cox proportional hazards models were fitted, treating smoking status and e-cigarette use status as time-dependent covariates updated at the date of follow-up health screening. (Paper page 12)
In time-varying exposure analyses restricted to participants with updated smoking and e-cigarette use information, precision was reduced and CIs [95% confidence intervals] widened, probably reflecting the reduced sample size and number of events. Nevertheless, the direction and trend toward a higher risk of lung cancer incidence among individuals who used e-cigarettes after smoking cessation were consistent (aHR 1.38, 95% CI 0.98–1.98; Supplementary Tables 7 and 8). (Paper page 2)
With regard to reverse causation, Kim and colleagues reported:
To address potential reverse causation, sensitivity analyses excluding cases that occurred within the first year of follow-up were conducted (paper page 12).
In sensitivity analyses excluding events that occurred during the first year of follow-up from baseline (Supplementary Tables 3–6), the associations between e-cigarette use after smoking cessation and the risks of lung cancer incidence and [lung cancer specific death] remained materially unchanged. (Paper page 2)
Rather than engaging this important information, Hajek just ignored it.
Indeed, I am still waiting for e-cigarette advocates to move beyond arguments based on what we knew (or assumed) 15-20 years ago and fully engage the current evidence base on e-cigarettes, including Kim and colleagues’ very well done paper.
And, of course, causality came up
Despite the fact that this was a huge well-controlled longitudinal study, the authors gave a nod to causality, the e-cigarette advocates and tobacco industry’s, last defense. The abstract ends softening its key conclusion, “these findings suggest that e-cigarette use after smoking cessation may attenuate the benefits of complete cessation for lung cancer prevention” by prefacing it with “[a]lthough causality cannot be established.”
This equivocating language was almost certainly added in response to Reviewers 2 and 4, who raised the issues of reverse causation (which the published paper addresses) and concern that “causality cannot be definitively established due to the observational design.” While the reviewer is correct that observational studies are not as strong as experiments in demonstrating causality, the fact is that it is impossible to do an experiment in which humans are randomly assigned to use or no use e-cigarettes then followed for years to see if they develop cancer (or any other disease).
(Given that “causality” has become something of a political whipping boy, I simply avoid the word in most of my papers. Of course, that hasn’t stopped pro-e-cigarette reviewers from criticising my papers as not proving causality. I respond by simply saying that I avoid the word “cause.”)
As Reviewer 4 notes, “Mechanistic experimental studies can be considered to explore causality and biological pathways.”
Kim and colleagues briefly discuss this broader evidence base, which was the subject on an extensive recent review. I would have been a lot happier if the nod to causality had not been made, since no one asserted that this one study, all by itself, would prove causality. But it sure provides strong evidence that, combined with the wide range of other evidence, e-cigarettes do cause lung cancer, especially because they only followed former smokers for 6 years. Given the long latency periods for cancer, it is likely that the cancer risks of e-cigarettes are even higher than this study found.
The dual use findings were surprising
Most studies of the association of dual use (using both e-cigarettes and cigarettes) is associated with increased disease risk, including lung cancer. In contrast, Kim and colleagues found that dual users had lower risks of lung cancer (incidence: aHR 0.78, 95% CI 0.72–0.85; death: aHR 0.66, 95% CI 0.55–0.79) and all-cause mortality (aHR 0.74, 95% CI 0.71–0.78) than current smokers at baseline who used only cigarettes. They suggested that this finding might be explained by the fact “that dual users were significantly younger and had significantly fewer cumulative pack-years of cigarette smoking than those who used cigarettes only.
As of now, however, most evidence indicates that dual use is more dangerous than just smoking. Figuring out why this study found such different results will be imnportant given its overall quality.
The bottom line
This is an exceptionally well-done strong study that seriously challenges the assumption that using e-cigarettes for smoking cessation is an effective harm reduction strategy. Quite the contrary, they increase harm in terms of lung cancer risk.
Here is the abstract:
Electronic cigarettes (e-cigarettes) have gained popularity as a less harmful alternative to conventional cigarettes, yet their associations with lung cancer risk after smoking cessation remain uncertain. Here we evaluated 4,524,895 adults with a conventional smoking history who participated in the Korean National Health Screening Program in 2018 (baseline), with prior records from 2012-2014. Participants were classified as current smokers, short-term quitters or long-term quitters, and followed up to December 2023. Daily e-cigarette use at baseline was used to define post-cessation e-cigarette use. Lung cancer incidence and lung cancer-specific death (LCSD) were assessed using multivariable Cox models. Over 24,182,543 person-years, 35,887 lung cancers and 12,807 LCSD events occurred. Compared with complete quitters, e-cigarette use after smoking cessation was associated with higher risks of lung cancer incidence (adjusted hazard ratio (aHR) 1.56, 95% confidence interval (CI) 1.24-1.97) and LCSD (aHR 2.00, 95% CI 1.28-3.15). Associations were directionally consistent in short-term and long-term quitters and were prominent in the high-risk subgroup (incidence: aHR 1.91, 95% CI 1.44-2.53; LCSD: aHR 1.92, 95% CI 1.13-3.24). Although causality cannot be established, these findings suggest that e-cigarette use after smoking cessation may attenuate the benefits of complete cessation for lung cancer prevention.
The full citation is: Kim YW, Park EJ, Kwak KI, Yoon TH, Kang HR, Joo DH, Lee YJ, Park JS, Lee JH, Lee CT. Electronic cigarette use after smoking cessation and lung cancer risk. Nat Med. 2026 Jun 8. doi: 10.1038/s41591-026-04469-5. Epub ahead of print. PMID: 42260103. It is available here.
Stanton Glantz blog 