Skip to main content
NCBI home page
The BMJ logoLink to The BMJ
. 2000 Aug 5;321(7257):378.

Smoking and dementia in male British doctors

Authors did not, strictly speaking, compare smokers with non-smokers

Eric Boyd 1
PMCID: PMC1118340  PMID: 10991556

Editor—Doll et al's finding that “persistent smoking does not substantially reduce the age specific onset rate of Alzheimer's disease or of dementia in general” is not surprising.1 The authors didn't compare smokers with non-smokers.

By combining lifelong non-smokers and ex-smokers in the non-continuing group they effectively stopped comparing smokers with non-smokers. To complicate the issue further they then note, “As questionnaires were sent out only every six to 12 years, the mean time before death that the relevant smoking habits had been recorded was not 10 but 15 years.” In the end this study compares a group including non-smokers and ex-smokers who may have started smoking in the previous 15 years with a group of smokers who may have stopped in the previous 15 years.

Has the BMJ fallen prey to the concerted and unrelenting efforts of health organisations determined to dictate an antismoking social policy rather than provide the honest and unbiased facts that people need to make informed personal choices? Or is the BMJ part of the team?

Footnotes

Competing interests: None declared.

References

  • 1.Doll R, Peto R, Boreham J, Sutherland I. Smoking and dementia in male British doctors: prospective study. BMJ. 2000;320:1097–1102. doi: 10.1136/bmj.320.7242.1097. . (22 April.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2000 Aug 5;321(7257):378.

Paper shows politically engaged research on smoking

Joao Calinas-Correia 1

Editor—Chalmers and Altman1-1 analysed the work of Janerich et al regarding passive smoking and lung cancer.1-2 Janerich et al focused on the effects on children, presenting a hypothesis that is contradicted by their own results (according to Chalmers and Altman), and failed to comment on the inverse relation they found between social exposure to passive smoking and lung cancer. This is just an example of the kind of bias likely when dealing with cigarette smoking in the current political environment.

Doll et al present another version of a politically engaged analysis.1-3 While they found that smoking seems to protect against Alzheimer's disease and eventually offers longlasting protection (the ex-smokers showed a decrease in risk similar to that in current smokers), their final comment reorganises the data to affirm the opposite. Although they present their results comparing current smokers with never smokers, ex-smokers with never smokers, and current smokers with ex-smokers, for their final comparison with other studies they create a different category—different from the categories they used in their discussion and different from the comparable studies. Is it prompted by their findings? No, it is not. They found a pattern where current and ex-smokers behave similarly, but they nevertheless aggregated ex-smokers and never smokers. With this trick, the difference they found between all smokers and never smokers was diluted and they could present the politically correct conclusion of no benefit.

Doll et al's results reinforce the evidence for protection against Alzheimer's disease, with a risk rate of 0.83 for continuing smokers and 0.78 for ex-smokers, refining, but fundamentally agreeing with, the results they quote from Herbert et al. The power to assert significance may be discussed, but the exercise in denying the trends they found is a sign of the submission of research to the political agenda.

Footnotes

Competing interests: None declared.

References

  • 1-1.Chalmers I, Altman D. Systematic reviews. London: BMJ Books; 1995. [Google Scholar]
  • 1-2.Janerich DT, Thompson WD, Varela LR, Greenwald P, Chorost S, Tucci C, et al. Lung cancer and exposure to tobacco smoke in the household. N Engl J Med. 1990;323:632–636. doi: 10.1056/NEJM199009063231003. [DOI] [PubMed] [Google Scholar]
  • 1-3.Doll R, Peto R, Boreham J, Sutherland I. Smoking and dementia in male British doctors: prospective study. BMJ. 2000;320:1097–1102. doi: 10.1136/bmj.320.7242.1097. . (22 April.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2000 Aug 5;321(7257):378.

Authors' reply

Richard Doll 1, Richard Peto 1

Editor—Our study found no association between the smoking habits about 15 years earlier and the likelihood of death from dementia (or, for those dying of other causes, the likelihood of dementia being mentioned on the death certificate).

Boyd says that our findings may have been distorted by changes in smoking habits during the long interval between the time when smoking habits were recorded and the deaths of doctors with dementia. We had made this interval long on purpose to ensure that the habits were recorded before the onset of appreciable disease, as illness may make smokers give up the habit, artificially inflating the proportion of recent ex-smokers among those who die with dementia. Few, however, of those who were not current smokers would have taken up the habit, to judge by later information on men of similar age who did not die and who replied to a subsequent questionnaire.

Our study involved a cohort in which only about 15% were lifelong non-smokers but in which half the smokers had given up the habit several decades ago. As there were fewer lifelong non-smokers than long term ex-smokers, our main assessment of whether persistent smoking affected dementia involved comparing all those who were still current smokers with all those who were not, three quarters of whom were ex-smokers who had given up an average of 34 years before death.

We found a relative risk of 0.96 (or 0.99, for dementia probably or definitely due to Alzheimer's disease), with no significant heterogeneity of risk between smokers, ex-smokers, and non-smokers. In our view, unduly selective emphasis just on the lifelong non-smokers would, in this context, be as inappropriate as unduly selective emphasis on the non-significant excess risk when continuing smokers are compared with long term ex-smokers.

Our overall conclusion that “persistent smoking does not reduce the age specific onset rate of the disease [Alzheimer's disease] or of dementia in general to any substantial extent” but “if anything . . . may increase rather than decrease the onset rate” was based not only on the evidence from our own study but on all that available from prospective studies. These other studies included two with relative risks for persistent smoking greater than 1.0 and not just the one such study with a (non-significantly) lower relative risk that Calinas-Correia chose to cite. Our conclusion therefore seems more soundly based than his.

Footnotes

Competing interests: None declared.

Articles from BMJ : British Medical Journal are provided here courtesy of BMJ Publishing Group

RESOURCES