PubMed2005
DATA SOURCES: Studies were located using Medline, Embase and the Cochrane Central Register of Controlled Clinical Trials, as well as searching bibliographies of identified references, review articles and consensus statements by hand. STUDY SELECTION: As the habit of smoking cannot be randomised, studies included were both controlled clinical trials and arms of randomised controlled trials investigating the effects of nonsurgical periodontal treatment, which reported results separately for smokers and nonsmokers. Other inclusion criteria were that studies should assess otherwise healthy people who had been diagnosed with chronic or adult periodontitis and that the patient was the unit of analysis (rather than a tooth- or site-based analysis). Studies were not excluded on the basis of quality, only on whether they fulfilled the inclusion criteria for entry. DATA EXTRACTION AND SYNTHESIS: General information about the paper, study characteristics, outcome measures, treatment characteristics and quality assessment was extracted independently, in duplicate. Where disagreement arose, this was resolved by discussion. Meta-analysis of data was performed and heterogeneity between studies was investigated using meta-analysis regression. RESULTS: Out of the 330 studies initially identified, 13 were considered eligible for inclusion. No studies reported data on tooth loss. For all sites, the reduction in probing depth (PD) was 0.13 mm greater in nonsmokers than in smokers (six studies) and there was no evidence to suggest that the studies were dissimilar in their estimates of this result (no evidence of heterogeneity; P>0.05). For sites with an initial PD of 5.00 mm (eight studies), a random-effects meta-analysis indicated a weighted mean difference in PD reduction of 0.43 mm favouring nonsmokers (95% confidence interval (CI), 0.24-0.63; P<0.001). Because of significant heterogeneity between studies, only a cautious observation can be made but, with one exception, all studies produced a summary estimate favouring nonsmokers. Meta-analysis of the two studies that compared the change in clinical attachment level between nonsmokers and ex-smokers, who had given up their habit, in sites with an initial PD of +/-5.00 mm, showed a difference in clinical attachment level gain of 1.34 mm favouring the nonsmokers (95% CI, 0.65-2.03; P<0.001; chi(2) test for heterogeneity, 7.47 with 1 degree of freedom; P=0.006). In both of these analyses, the degree of heterogeneity is a cause for concern. Bleeding was assessed after therapy in seven studies but meta-analysis was not performed because of great heterogeneity in the methods used to assess bleeding. No statistically significant differences in bleeding were found between smokers and nonsmokers either at baseline or after therapy in most of the studies. One study found significantly less bleeding in smokers than in nonsmokers at baseline and another found a reduced response in terms of bleeding in smokers than in nonsmokers. Two studies evaluating the change in bleeding in ex-smokers found no statistically significant difference after treatment. No data were reported for any of the included studies on patient-centred outcomes such as quality of life, ease of maintenance, changes in aesthetic appearance, or patient experience. CONCLUSIONS: Following nonsurgical periodontal therapy, people who smoke will experience less reduction in PD than nonsmokers. There is no evidence of a difference in gain in clinical attachment between smokers and nonsmokers or a reduction of bleeding on probing between smokers and nonsmokers. Differences in study design and lack of data precluded an adequate and complete pooling of data for a more comprehensive analysis. In short-term studies, it is unclear whether people who stop smoking will respond as favourably to nonsurgical therapy as those who have never been smokers. Progress in understanding the effects of smoking on periodontal therapy will require the evaluation of objective measures of smoking such as nicotine exposure and exhaled carbon monoxide in place of sole reliance on patient-reported information.