Commentary

The systematic review of Hanioka and colleagues is a landmark report. The authors comprehensively, competently, and carefully identify, appraise, and review observational studies linking smoking to tooth loss. They then convincingly demonstrate a substantial significant causal relationship between smoking and tooth loss.

These results caused me to question my clinical, legal and ethical responsibilities in pursuing caries and periodontal disease control among smokers, without simultaneously addressing smoking cessation. After all, the clinical measures we use for both caries and periodontal disease are really surrogates for the ultimate measure – tooth loss. If a patient comes to see me to 'save my teeth', am I also responsible for smoking cessation, or at least appropriate informed consent and referral? And what is appropriate informed consent?

Let's be more specific. The systematic review indicates that the odds of losing teeth are two to four times higher in smokers than non-smokers. The converse suggests that I will be successful in saving teeth only half to a quarter of the time in smokers. And then I dived into the report. The authors searched three databases for observational trials and identified 15 trials: five cohort and ten cross-sectional, from five countries. They appraised the trials for quality (absence of bias) using the Newcastle-Ottawa scale. Eight studies were categorised as high quality – two cohort and six cross-sectional, in four countries – and the remainder were categorised as moderate quality. Using the results of the included studies, the authors categorised their findings according to the Bradford Hill criteria, and in every case the results met these criteria.

  • Consistency of findings. All six studies provided similar directionality; that is, smoking is associated with greater tooth loss.

  • Strength of association (magnitude and statistical strength). Among current smokers the odds ratio ranged from 1.7 to 4.0. That is, for smokers the odds were 2-4 times higher that they would be missing teeth than non-smokers, and all results were statistically significant.

  • Biological gradient (dose response relationship). Four studies examined cigarette “consumption” and tooth loss among current and former smokers, and all identified a similar trend: increased smoking is associated with increase tooth loss.

  • Natural experiments. One might expect former smokers to have reduced tooth loss compared with current smokers. This too was found to be true.

  • Biological plausibility. From studies other than those reported here the authors identify the proposed mechanisms by which the components of tobacco smoke can generate inflammatory damage and alter the oral microbiology from health to disease-related.

Hanioka et al. were, to my mind, overly self-critical in underscoring the limitations of their report. While there is certainly no perfect study, and no perfect systematic review, the authors deserve substantial commendation for the yeoman's task they took on in dissecting out the nuances of the reports and collating them into a very usable and important whole. Their work truly establishes a base for professional self-assessment in terms of how we address smoking.