select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

ecig vapour and cancer in mice

Research, published in PNAS, reports that e-cigarette vapour may be carcinogenic in the lungs of mice. 

A Roundup of comments accompanies this Before the Headlines. 



Title, Date of Publication & Journal

‘Electronic-cigarette smoke induces lung adenocarcinoma and bladder urothelial hyperplasia in mice.’ Proceedings of the National Academy of Sciences of the USA.


Study’s main claims – and are they supported by the data

The study’s main claim is that electronic cigarette (ecig) vapour causes lung adenocarcinoma (lung cancer) and bladder urothelial hyperplasia (an increased number of cells lining the bladder) in mice.

This is somewhat supported by the data in the study, though there are a number of serious limitations in the study which may have caused considerable bias, therefore affecting the validity of the data and subsequently this claim.

It would be wrong to conclude from this study that vaping is carcinogenic in humans. Notably, the authors do not state that their study is conclusive – only that it warrants further research, which is a balanced conclusion to make and is something it would be difficult to disagree with.




  • The study is a randomised trial and so therefore has the potential to demonstrate a causal relationship in mice; however, there are a number of limitations which could have biased the study’s results (see below).
  • The authors have used ecig juice and aerosol generators that are typically available commercially.
  • Although there is a huge amount of uncertainty in the results, there is the potential for a fairly large increased risk in adenocarcinoma in mice exposed to ecig vapour.


Limitations of the study

  • Was this a proper blind trial? It is not clear whether the group allocation was kept secret from either the researchers administering the compounds or those assessing the histopathology. This is known to potentially lead to very biased results.
  • The statistical evidence of an effect on lung carcinoma is very weak
  • The authors perform multiple statistical tests which increases the probability of a chance positive finding
  • The choice of the statistical test is not appropriate for this type of data.  Generally the stats in this paper are very poor.
  • The authors do not provide indicators of uncertainty (such as confidence intervals) for all the comparisons, which makes impossible to assess the precision of the results.
  • The study has a small sample size and they have not performed a power analysis (see glossary)
  • The authors say animals were randomly allocated to different groups, but the numbers in each group (45, 20, 20) are perfectly rounded, which would be very unlikely to occur by chance without using a more technical method or randomisation. This may suggest the authors have not used a truly random technique for allocation, which is known to introduce a lot of bias. Moreover, mice dying before the end of the study were excluded from the analysis, which may also cause bias.

A final point is that this study is in mice, not in humans, and mice may respond to nicotine or ecig vapour differently for a number of reasons.

It is also unclear if the exposure on the mice is typical of exposure in humans.



Power analysis: A statistical calculation done before running an experiment to ensure you use a large enough sample size to be able to identify the result you are looking for with sufficient precision.


Any specific expertise relevant to studied paper (beyond statistical)?



Before The Headlines is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research.  A list of contributors, including affiliations, is available at

in this section

filter Headlines by year

search by tag