Abstract
Selection Criteria
Two investigators independently searched 3 databases (PubMed, Web of Science, and Embase) with hand search of the reference list from retrieved articles. Search was restricted to studies published in English up to July 2015. A third investigator arbitrated when there were differences in opinion regarding the eligibility of a study. Inclusion criteria were observational studies published as original articles with the purpose of evaluating the association between alcohol consumption and periodontitis, and studies that report multivariate-adjusted relative risk (RR) or odds ratio estimates with 95% confidence intervals (95% cls) or included data that enabled their calculation. For dose-response analysis, the studies providing RRs or odds ratios of periodontitis as well as number of cases and person-years for at least 3 quantitative categories of alcohol consumption were included. In cases of more than one study including the same study population, the data from the most recent study or the one with the largest sample size were included.
Key Study Factor
The key study factor was alcohol intake. Nondrinking/occasional drinking was used as the reference category. The authors performed a highest vs lowest intake analysis and a dose-response analysis. The highest vs lowest analysis compared the highest alcohol category in a particular study, regardless of the actual dose (eg, >= 30 g ethanol/d, >= 10 drinks/wk), to the reference group.
Main Outcome Measure
The primary outcome measure was periodontitis defined by various diagnostic criteria including clinical attachment loss (CAL), pocket probing depth (PPD), community periodontal index, and alveolar bone loss. Clinical diagnostic, radiographic diagnostic, or both methods to assess CAL were included. Pooled RRs of periodontitis comparing highest and lowest intake were computed, with corresponding 95% Cls. Subgroup analyses were performed by the type of study design, continents where the studies were performed, and diagnostic criteria of periodontitis. A sensitivity analysis was performed to evaluate the key studies that have substantial impacts on between-study heterogeneity. All results were reported as RRs regardless of study design or the effect estimate stated in the original study.
Main Results
Highest vs lowest analysis revealed that 11 studies of 18 showed a statistically significant association between alcohol intake and periodontitis. The pooled RR for the highest vs the lowest alcohol consumption level was 1.59 (95% CI: 1.37-1.85). When stratified by gender, the risk of periodontitis with high alcohol use was doubled among females but only 25% higher in males. Stratification by continent revealed that periodontitis risk ranged from a 45% increase in studies from Asia to a 76% increase in studies conducted in America and Europe. In the subgroup analysis by study design, cross-sectional studies showed a larger effect (66% increase) than cohort studies (28% increase). In the subgroup analysis by diagnostic criteria of periodontitis, PPD and PPD plus CAL showed the largest effects of alcohol intake. The dose-response analysis of alcohol consumption and periodontitis risk showed a linear relationship. Overall, the risk for periodontitis increased by 0.4% (P = .002) for each 1 g/d increment in alcohol consumption. When analyzed by intake category, the RRs were statistically significant only at intake levels >= 30 g/d.
Conclusions
The authors concluded that alcohol consumption was associated with an increased risk of periodontitis and that a linear dose-response relationship existed. They recommended that alcohol consumption be recorded to estimate the behavioral risk for patients.