This meta-analysis reveals association between PD and increased risk of severe disease following COVID-19 infection, which is consistent with previous studies reported individuals with PD tend to be more negatively affected by COVID-1930. It is noteworthy that results of the current meta-analysis regarding risk of severe disease should be interpreted with caution because different outcomes of included studies were pooled for the main outcome of severe disease. In addition, our analysis included studies with relatively small sample sizes. Thus, a significant heterogeneity was observed. Studies have shown that ACE-2 is the main receptor for SARS-CoV-2 to enter human cells. ACE-2 mainly exists in lung, nasopharyngeal mucosa, salivary gland and oral mucosa.4,31 Flynase, a protease in the oral cavity that is elevated in chronic periodontitis, can prefragment the coronavirus spike protein (S protein) into two subunits (S1 and S2), and the S1 receptor binding domain (RBD) can attach itself to ACE-2. Finally, upon binding of the S1 subunit to the ACE-2, the virus passes through the host cell to fuse and form a nucleus, resulting in the closure of the cell membrane of the virus and the host and leading to cell fusion and infection.32 Therefore, it can be hypothesized that increased protease levels in periodontitis may increase the risk of oral mucosa-mediated SARS-CoV-2 infection, and even contribute to worsening COVID-19 infection. In addition, people with PD may have an increased risk for subsequent systemic diseases, including cardiovascular disease, hypertension, respiratory disease, and diabetes.33–37 This might contribute to development of poor outcomes if PD patients are infected with SARS-CoV-2. Besides, it has been claimed that cytokines produced during periodontitis can promote adhesion of respiratory pathogens to the lung epithelium and colonisation of the lungs, and inhaled periodontal bacteria may induce cytokine secretion in the lower respiratory tract, exacerbating the symptoms of COVID-19.38,39
There was no increase in hospital admission among COVID-19 patients with PD compared to those with healthy periodontal condition, which contradicted conclusions of Gupta et al.5 This may be related to the fact that the included studies had different diagnostic criteria for exposure, e.g. Larvin et al.12 used self-diagnosed periodontal disease while other studies used rigorous examination methods. Compared with COVID-19 cases without PD, those with PD had a 0.61-fold increased risk of ICU admission, a 1.25-fold increased risk of death, and a 4.01- and 4.19-fold increased risk of mechanical ventilation and pneumonia, respectively. Periodontitis has long been thought to have its pathophysiological roots in cytokine responses.40 In addition, the adverse outcomes of the COVID-19 infection have been reported to be associated with cytokine storms. These adverse outcomes also share some similar cytokine expression profiles to periodontitis.41 Several studies also demonstrated that, in patients with gingivitis, the number of cytokine and chemokine producing cells in gingival tissues and cytokine level in serum increased significantly comparing to healthy controls.42 Two studies have shown that patients requiring assisted ventilation presented higher CRP and D-dimer blood levels.5,10 This is consistent with previous studies reporting elevated inflammatory markers in patients with COVID-19 who passed away.43 CRP levels are considered to be an early biomarker of the severity of COVID-19 infection, and elevated D-dimer has been reported as a prognostic marker associated with a worsening prognosis in patients with COVID-19.44,45
Subgroup analyses based on study type were performed, and statistical associations of PD with COVID-19 severe disease were found in case-control study and cohort study but not in cross-sectional study. In the subgroup analysis based on study location, this study found a statistically significant association in the Asia, whereas not in European and other location. Differences in COVID-19 administration policy, vaccination schedule, and prevalent strains may also have contributed to this variability. In this study, a subgroup analysis was performed based on the original study quality assessment scores. Both groups of studies showed that PD was associated with the development of COVID-19 to severe disease. Meta-regression was used to explore the sources of heterogeneity, and the results concluded that both COVID-19 outcomes and study design type had an effect on heterogeneity, whereas region and study quality were unlikely to be sources of heterogeneity.
Notably, to our knowledge, this is not the first meta-analysis to examine the relationship between PD and COVID-19. In contrast to the previously published study by Baima et al,46 we chose observational studies for the meta-analysis and included a larger number of studies. In addition, there were some differences between our findings and the results of the former meta-analysis, the results of the previous meta-analysis found that periodontitis was associated with increased odds of hospitalization, however, this association was not found in our study. Furthermore, in a subgroup analysis of COVID-19 outcomes, we added the outcome of pneumonia complications and found a significant association. Finally, we conducted more detailed subgroup analyses based on study type, study region, and quality assessment scores to further explore sources of heterogeneity.
There are several inherent limitations to our study. Firstly, this meta-analysis included studies with relatively small sample size, and in terms of COVID-19 patient outcomes, two studies included only pneumonia, one study only analyzed death, and one study only explored the risk of assisted ventilation, therefore the results may be biased. Secondly, there was a high degree of heterogeneity in the overall analysis, which may have been due to variations in types of studies included. During the pandemic, some COVID-19 patients may not be able to receive periodontal examinations, and self-reported periodontal bleeding and pain may be associated with other dental conditions such as dental caries. Third, the lack of raw data on gender, age, comorbidities, and vaccination status prevented appropriate subgroup analyses to explore these possible sources of heterogeneity. Last but not least, Publication bias was detected in our study. Despite these limitations, the present study shows the following strengths. This is the first meta-analysis of observational studies describing the relationship between PD and COVID-19 with a larger number of subjects, and the vast majority of the studies we included adjusted for the same confounders such as gender, age. Therefore, the results were more reasonable, stable, and convincing. Secondly, different effect models are used according to the level of heterogeneity and the study results are more rigorous and credible. Thirdly, the sensitivity analysis found a robust result, prompting our results to be credible.