Since 2012, the causal link between periodontal disease and the incidence of malignant neoplasms, particularly oral and oropharyngeal cancers, has been a popular topic. Correlations between periodontal disease and the risk of head and neck cancers, particularly oral and oropharyngeal cancers, have been widely studied, with most studies reporting positive correlations, including studies from the United States [42], Europe [43], Latin America [44], India [45], China [46], and Brazil [47]. Correlation studies on periodontitis and pre-cancerous oral lesions (leukoplakia) also reported positive results [48, 49]. In contrast, we did not identify a causal relationship between acute or chronic periodontitis and oral cancer.
A clinical study that included 13,798 adults reported an association between periodontal disease severity and an increased overall risk of oral tumours after adjusting for several confounders, including smoking and alcohol consumption (OR = 4.57, 95% CI = 2.25–9.30) [48]. In addition, Rezende et al. [50] reported a positive correlation between advanced periodontal disease and oral cancer in a study including 35 patients and 40 healthy participants, irrespective of the patients’ oral hygiene and dental status. Another case-control study involving 200 oral/oropharyngeal cancer cases and 200 healthy participants reported a significant correlation between the number of missing teeth and oral/oropharyngeal cancer (≥ 16 teeth; OR = 2.74, 95% CI = 1.23–6.12) [51].
Moreover, several meta-analyses have evaluated the association between periodontal disease and head and neck tumours, also yielding positive results. A meta-analysis by Zeng et al. [52] included results from two cohort studies and six case-control studies, reporting a significant correlation between periodontal disease and the risk of head and neck cancer (OR = 2.63, 95% CI = 1.68–4.14). Wang et al. [53] also conducted a meta-analysis that included eight case-control studies and one cross-sectional study (n = 5,204 cases, n = 5,518 healthy participants), finding that tooth loss was significantly associated with developing head and neck cancer (OR = 2.00, 95% CI = 1.28–3.14), and this trend was more pronounced in those with moderate (18%) and severe (54%) tooth loss. Similarly, a meta-analysis by Zeng et al. [54] included X cohort studies and ten case-control studies found a significantly increased risk of head and neck cancer among participants who lost 6–15 teeth (OR = 1.58, 95% CI = 1.08–2.32), and meta-analyses by Yao et al. [55] and Ye et al. [56] also indicated an association between periodontal disease and oral cancer risk.
However, some studies reported mixed or negative results. For instance, a case-control study of a Carolina population examined 1,361 cases and 1,289 frequency-matched controls, finding that a self-reported history of loose teeth was positively associated with head and neck squamous cell carcinoma after controlling for possible confounders (OR = 1.33, 95% CI = 1.07–1.65). However, this correlation was not significant among never-smokers. They also reported that the degree of tooth loss (16–28 vs 0–5) was not associated with head and neck squamous cell carcinoma risk [57]. Furthermore, a large cohort study reported no significant association between periodontal disease and oropharyngeal cancer in adult male subjects (OR = 1.15, 95% CI = 0.73–1.81) [58]. In addition, a meta-analysis of seven case-control studies reported a linear relationship between the number of missing teeth and the risk of oropharyngeal cancer. For each additional missing tooth, the ratio significantly increased by 0.03 (95% CI = 0.01–0.05) [59]. However, a linear dose effect was not observed for head and neck cancer risk. A Taiwan-based cohort study (n = 148,166) reported a positive association between chronic periodontitis and oral cancer (OR = 1.20, 95% CI = 1.09–1.33) [60], and another cohort study of older women in the United States (n = 65,869) found no association between self-reported periodontal disease and lip, oral, and pharyngeal cancers (OR = 1.10, 95% CI = 0.64 − 1.87) [61].
Study-related differences might explain these contradictory findings. For example, some studies controlled for the effects of risk factors, such as smoking and alcohol consumption, but others did not. Different evaluation indices for periodontal disease were used among studies, and many of the studies included a small sample size, especially regarding the number of patients with head and neck (especially oral cancer) tumours. Finally, most studies were case-control studies, making it difficult to establish a causal relationship.
MR studies have been widely used to evaluate the causal relationships between risk factors and disease outcomes using genetic variants as instrumental variables [62]. To date, this is the first TSMR study to explore the causal relationship between periodontal disease and oral cancer. We used a large-scale GWAS database to identify 11 SNPs associated with acute periodontitis and oral cancer and 17 SNPs associated with chronic periodontitis and oral cancer in a European population. Then, we analysed the causal relationship between the two using five complementary MR methods, finding no causal relationship between acute or chronic periodontitis and oral cancer. Additionally, acute and chronic periodontitis were not risk factors for developing oral cancer. Therefore, causal relationships did not exist, and acute and chronic periodontitis were not risk factors for oral cancer.
This study has some limitations. First, the TSMR method assumes that the correlation between the exposure factor (acute/chronic periodontitis) and disease outcome (oral cancer) is linear; thus, this method is not applicable if the relationship is non-linear. Second, analysing statistical data in a database stratified by sex or age is difficult, potentially leading to biased results. Nonetheless, this study has several important advantages. This study used the “TwoSampleMR” R package to integrate the screened SNPs loci, exclude SNPs directly related to oral cancer risk factors, and fully consider and eliminate SNPs that highly affect disease outcomes. Second, stringent quality control conditions and analytical methods were adopted, causal effects were explored using five complementary MR analysis methods, and the robustness of the results was verified using two sensitivity analyses. Finally, using MR methods minimises the influence of confounding factors or reverse causality on the study results.