Association between the severity of periodontitis, COVID-19, C-reactive protein and interleukin-6 levels in hospitalized patients: a case‒control study

Hospitalized patients with a confirmed COVID-19 diagnosis who matched the inclusion criteria (explained in the next sections) were enrolled in the study. During the hospitalization of each patient, a detailed periodontal examination was done (discussed in the “ Periodontal Examination” section), and a sample was taken from their serum and saliva (discussed in the “Serum and saliva sampling” section). After the end of the hospitalization, medical data were collected, and the COVID-19 severity assessment was done. Patients who matched the criteria for mild to moderate COVID-19 were considered controls, while patients who matched the criteria for severe to critical COVID-19 were considered the case group. Patients were then retrospectively investigated for the evaluation of the hypotheses.

The distribution of periodontal disease severity and CRP and IL-6 levels among the case and control groups was retrospectively evaluated to investigate any possible association between COVID-19 and periodontitis (i.e., the primary exposure variable was periodontitis severity and the primary outcome variable was COVID-19 severity). The primary objective was to evaluate the association between periodontitis severity, COVID-19 severity, and levels of IL-6 and CRP. Secondary objectives included evaluating the associations between background characteristics and disease severity, between periodontitis severity and hospitalization length, and between periodontal parameters and adverse events. Analyses of hematological tests, such as complete blood count, electrolytes, coagulation, and other tests, are performed in the second part of this study [28]. This study follows the STROBE guidelines for reporting its results.

This part of the study shares its setting with the other part [28]. The details of the study setting were the following: Enrollment and examinations were done at a hospital with dedicated COVID-19 ward and intensive care unit (ICU). Patient enrollment was done by a single researcher at each hospital visit, which were scheduled randomly. At each hospital visit, all patient files in the COVID-19 ward were screened by this researcher, and patients who had an established COVID-19 diagnosis and matched the inclusion criteria were approached for written informed consent and subsequently enrolled in the study. Patients were assured that the data associated with them would be anonymized and could not be used to identify them, they would remain anonymous, and their COVID-19 treatment course would not be impacted by study participation. An extensive periodontal examination, in addition to saliva and serum sampling, was done for each enrolled patient, and they were followed for changes in their COVID-19 severity during their hospitalization.

Patient enrollment was done from December 2021 to October 2022. The study population consisted of hospitalized patients with an established definitive COVID-19 diagnosis (positive SARS-CoV-2 PCR test and clinical COVID-19 symptoms). This part of the study shares its inclusion and exclusion criteria with the other part of the study:

A single periodontist was responsible for all periodontal examinations. Periodontal examination was performed using a mirror and a standard periodontal probe with William’s markings (Hu-Friedy, Chicago, Illinois, US). Three periodontal indices were measured: the clinical attachment level (CAL), the modified sulcus bleeding index (MSBI), and the gingival index (GI). The number of remaining teeth (excluding third molars) was also recorded.

To calculate the CAL, the distance from the base of the periodontal pocket to the cementoenamel junction (CEJ) was measured using a light (0.25 N) probing force. The criteria originally proposed by Löe & Silness [29] were used to calculate the GI: 0 = normal gingiva without inflammation; 1 = mild inflammation, slight color changes, and slight edema; 2 = moderate inflammation, redness, edema, and bleeding on probing; 3 = severe inflammation, marked erythema, ulceration, and spontaneous bleeding. The MSBI was used as a measure to quantify the degree of bleeding on probing and was calculated using the criteria proposed by Mombelli et al. [30]: 0 = no bleeding; 1 = isolated points of bleeding; 2 = a confluent line of bleeding on the gingival margin; and 3 = heavy or profuse bleeding. These indices were measured around every tooth (excluding third molars) at six sites (mesiobuccal, mid-buccal, distobuccal, distolingual, mid-lingual, and mesiolingual). The highest measurement for each index in each sextant was recorded.

Initially, 142 patients were included in the study. Twenty patients were excluded from the study due to various reasons, upon the conclusion of follow-ups and a thorough examination of medical records. The diagram for study enrollment and the reasons for exclusion are presented in Fig. 1.

The mean, minimum, and maximum for age were 52.82 ± 13.34, 19, and 78 years, respectively, exhibiting a normal distribution. Of all patients, 48.4% (n = 59) were female. The minimum and maximum oxygen saturation with passive oxygenation for the entire hospitalization course were 86.16 ± 7.2% and 93.33 ± 3.71%, respectively. The most frequent appearance of lung involvement in CT scans was ground glass opacities, alone (n = 63, 52.9%) or combined with other appearances. Hypertension (33.6%, n = 41), diabetes mellitus (13.1%, n = 16), and chronic obstructive pulmonary disease (10.7%, n = 13) were the most common findings in medical histories. Of the patients with available smoking histories (n = 111), 70.3% (n = 78) had no smoking history. Nine (7.4%) patients had respiratory distress initially or during their hospitalization, and 27 (22.1%) of the patients required intensive care during their hospitalization. The liver was the most prevalent (13.9% of all patients, n = 17) organ with COVID-19 associated involvement. Adverse events occurred in 27.9% (n = 34) of patients, including three deaths. Length of stay (LOS) did not have a normal distribution, and the mean, median, and interquartile range for it were 10.24, 8, and 5 days, respectively. The general results can be found in Table 1.

There was a significant age difference between the COVID-19 and periodontitis groups. Independent t-test revealed that the SC group had a significantly higher average age than the MC group (P = 0.004). One-way ANOVA revealed a significant age difference between the three periodontitis groups (P < 0.001). Chi-square tests indicated no significant difference regarding sex between COVID-19 groups (P = 0.69), while there was a significant difference regarding sex between periodontitis groups (P = 0.01, Chi2 = 9.17). The results for these tests can be found in Table 2.

To test for a possible association between COVID-19 and periodontitis, the chi-square test was used. The results indicated that the severity of periodontitis and COVID-19 were significantly associated (P < 0.001). In fact, 74.4% (n = 32) of the SP group had severe to critical COVID-19. The results can be found in Table 3.

To obtain odds ratios for periodontitis and COVID-19, while accounting for IL-6 and CRP levels and controlling for confounders, logistic regression was used. Compared to the HP group, the SP group had an adjusted odds ratio (AOR) of 4.24 for severe to critical COVID-19 (P = 0.047, 95% confidence interval: 1.020 – 17.683). IL-6 had significant association with severe COVID-19. Each unit increase in serum IL-6 levels would increase the odds of severe COVID-19 by 1.075 (P = 0.033, 95% confidence interval: 1.006 – 1.148). The results for crude odds ratios and other details can be found in Table 4.

The ANCOVA adjusted for age, sex, and comorbidities revealed that none of the single nor combined models could account for changes seen in CRP levels across groups. On the other hand, the single models revealed that salivary and serum IL-6 levels had a significant association with COVID-19 (P values: < 0.001 and < 0.001, respectively) and periodontitis severity (P values: 0.002 and 0.006, respectively). The combined models revealed that while the interaction between periodontitis and COVID-19 did not have significant effect on IL-6 levels, the main effect of COVID-19 was significantly associated with serum and salivary IL-6 levels (P values: 0.002 and 0.004, respectively). These results can be found in Table 5.

A significant association between CAL and adverse events (P = 0.004) and GI and adverse events (P = 0.035) was found using logistic regression. It was also revealed that the number of remaining teeth was inversely associated with adverse events (P = 0.023). No significant association was found regarding MSBI. The results are shown in Table 6.

To investigate the association between LOS and periodontitis, the Kruskal‒Wallis test was used. The results showed that periodontitis was significantly associated with LOS (P = 0.004). The results are shown in Table 7.

In this study, we investigated the association between the severity of COVID-19 and the severity of periodontitis in hospitalized patients. We also investigated the possible links between these two diseases and other factors, such as age and sex. Age had a significant association with both COVID-19 and periodontitis severity. The SC group was approximately 7 years older than the MC group. This finding was similar to other studies [25, 26, 38, 39]. The explanation for this association is mainly a phenomenon called immunosenescence (i.e., immune system aging), which entails an increase in memory T cells and a decrease in naïve T cells. These cells also show less potential for proliferation but more potential for proinflammatory cytokine production [40]. More severe forms of periodontitis were also associated with increased age. The mean age of the HP group was 43.16 years, while that of the SP group was 59.02 years. The current consensus is that periodontitis is an age-associated disease, with an abundance of literature reporting similar findings [20, 41‐44]. The reasons for this association are not yet entirely understood, but changes in innate and adaptive immunity have been at the forefront of possible causes [44].

Although there were more males in the SC group (53.6%, n = 30), the association between sex and COVID-19 severity did not reach statistical significance. Many studies have found that males are more predisposed to severe forms of COVID-19 [26, 45, 46]. Estrogen-mediated upregulation of ACE2 [47], which protects against the damaging effects of SARS-CoV-2-induced ACE2 downregulation, decreases the odds of females progressing toward more severe forms of COVID-19 [48]. Similarly, it has been documented that periodontitis is more prevalent in males [49], which is in line with our findings. A total of 47.6% (n = 30) of males were in the SP group, while only 22% (n = 13) of females were in the SP group. Reasons for this disparity are categorized into two major groups: biological differences between sexes (e.g., immune system, etc.), and social differences (e.g., socioeconomic factors, cultural attitudes, access to preventive and regular care, etc.) [50].

In this study, the association between the severity of periodontitis and COVID-19 was found to be statistically significant. This finding was similar to the study performed by Anand et al., which showed that severe forms of periodontitis had an odds ratio of 11.75 for COVID-19 infection [51]. Two other studies also found that severe forms of periodontitis were associated with higher odds of adverse events and complications and, hence, more severe forms of COVID-19 [52, 53]. We found that more severe forms of periodontitis had higher odds for severe COVID-19. The SP group had an adjusted odds ratio of 4.24 for severe COVID-19, compared to the HP group. Marouf et al. reported an AOR of 3.67 for adverse events, 3.54 for ICU admission, and 8.81 for death [52]. In agreement with the above findings, Gupta et al. reported an AOR of 7.45 for mechanical ventilation, 36.52 for hospital admission, and 14.58 for death in severe forms of periodontitis, compared to healthy patients [53]. Regarding periodontal parameters, we found significant associations between CAL and adverse events, and between GI and adverse events. Similar results were also found by Gupta et al. and Anand et al. [51, 53]; Gupta et al. found significant associations between bleeding on probing, mean probing depth and COVID-19 severity indicators such as hospitalization, need for assisted ventilation, and survival. Anand et al. found significant differences in pocket depth, CAL, and gingivitis between their case and control groups. It can be said that due to increases in systemic inflammation levels caused by severe forms of periodontitis, the COVID-19 course will be more likely to progress to a hyperinflammatory state and eventual cytokine storm. Subsequent to the cytokine storm, organ failure and medical complications [12] will necessitate intensive care and treatments in order to sustain the patient’s life.

Interleukin-6 is a major proinflammatory cytokine and one of the main signaling molecules responsible for the production of acute-phase proteins such as CRP [54]. It has been found that higher levels of this cytokine adversely affect the survival of COVID-19 patients [25, 26]. Compared to individuals with a healthy periodontium, patients with periodontitis show elevated levels of IL-6 [22, 24, 42]. In this study, we found a significant association between COVID-19 severity and IL-6 levels. C-reactive protein is a nonspecific acute-phase protein produced in the liver that is upregulated by inflammatory processes and may be an indicator of systemic inflammation and the severity of COVID-19 [55, 56]. In regards to periodontitis, it has been well documented that higher CRP levels are found in the saliva and serum of periodontitis patients compared to healthy individuals, and following treatment for periodontitis, CRP levels decrease [20, 22, 41]. The CRP increase in periodontitis is due to the fact that the chronic inflammation process of this disease increases systemic pro-inflammatory cytokines [20, 24], which would lead to the upregulation of inflammation markers such as CRP [57]. In our study, despite increased CRP levels in more severe forms of periodontitis and COVID-19, none of the adjusted statistical models could meaningfully account for the observed differences. In a review article, most of the reviewed studies reported increased CRP levels in severe forms of COVID-19 [56]. Higher levels of this biomarker are not only an indicator of the severity of COVID-19 but also an indicator of the prognosis of patients [26, 58].

IL-6 and CRP levels and their associations should be interpreted with caution. The most critical concern is that the timing of the serum and salivary sampling was not completely standardized. Even though all samples were collected between 9 and 11 a.m., one patient could be sampled on the 3rd hospitalization day, while another could be sampled on the 8th day. Coupled with the fact that the interval between the onset of symptoms and hospitalization was different between patients, this would mean different inflammatory timepoints in the disease course, resulting in different cytokine and marker profiles [59]. Another concern was the treatment regimens/medications for patients, especially immunomodulatory and immunosuppressive drugs. Although many patients received similar principal medications, and there was no significant difference regarding medications across groups, the dosages and response could have been different between patients, which may result in varying levels of host response, such as immune system activity. Furthermore, it has been established that there are genetic variations that may predispose individuals to higher risks for developing certain diseases or higher risks for progressing toward more severe conditions. In a recent meta-analysis by da Silva et al., two interleukins whose genetic variances could significantly increase the risk for chronic periodontitis were identified: IL-1α and IL-1β [60]. Specific genetic mutations of IL-1α were investigated in another review study and found to increase the severity of COVID-19 [61]. The IL-10 mutation rs1800872 has been found to increase the susceptibility of individuals to both periodontitis and COVID-19 [61‐63]. These findings and future findings may further reveal the role of genetic variations in disease progression and may indicate that serum levels of interleukins alone may not be able to completely explain all findings.

The mean length of stay for all patients was 10.24 days. The mean LOS for the MC and SC groups were 7.06 and 13.98 days, respectively. The mean LOS for the HP, MP, and SP groups were 8, 9.64, and 12.74 days, respectively. The increase in LOS was significantly associated with the severity of periodontitis. There were no previous studies investigating the association between periodontitis and LOS in COVID-19 patients, but some studies have investigated the association between COVID-19 and LOS [39, 46, 64]. The main reason for this finding is postulated to be inadequate viral clearance, which at the time of seroconversion may lead to antibody-dependent enhancement and further worsening of the clinical course [8]. Periodontitis may be associated with hospitalization length due to the direct and indirect effects it has on COVID-19 severity.

There are many underlying pathways that link periodontitis with COVID-19, however, only a few have been mentioned thus far. While the genetic differences and cytokine connection have already been covered, other pathways also need to be taken into consideration. It has been demonstrated that SARS-CoV-2 entrance receptors, including ACE2 and CD147, are present in the oral cavity, particularly on epithelial cells [6, 7, 65], and therefore in the periodontium. It has also been shown that SARS-CoV-2 is present in the gingival sulcus [19] in detectable numbers high enough to be utilized as a reliable diagnostic test. Furthermore, there have been findings regarding the direct pathosis caused by SARS-CoV-2 in the periodontium, including direct fibroblast degeneration and senescence, hyperproliferation, and pathogenic fibrotic phenotypes [66]. Periodontitis can affect COVID-19 as well, with emerging evidence showing an association between the two diseases [51‐53]. The elevation of systemic inflammation and the increase in cytokine levels in more severe forms of periodontitis [22, 24, 42, 67], many of which are directly associated with COVID-19 progression and adverse outcomes [26, 68, 69] is perhaps the most obvious way in which periodontitis can affect COVID-19. The second potential link might be between periodontitis and systemic comorbidities [70]. The effect of periodontitis on these comorbidities might indirectly influence COVID-19 infection due to the significant effects these comorbidities have on COVID-19 [71], many of which are already known [13, 46]. Finally, SARS-CoV-2 can independently replicate in the oral cavity and periodontium, establishing a reservoir in this region. The virus can then spread through the body via hematologic, digestive, or respiratory routes [71]. More information about the underlying pathways linking periodontitis and COVID-19 will become clear as more studies are conducted.

A clinical diagnosis of periodontitis was made in this study, which is one of its strengths; nevertheless, one of its weaknesses was the lack of radiographic information to support the diagnosis. A hospital setting allowed us to study the more severe forms of COVID-19 in a more controlled environment. The retrospective assessment of COVID-19 severity might be prone to inaccuracies or inconsistencies in the documentation. We attempted to mitigate this issue by evaluating all available records during the entire hospitalization course, whether paper or electronic. As discussed previously, the sampling day was not standardized between patients, which might be a limitation for this study.

Studying hospitalized patients served as a double-edged sword. While it provided us with the means to study the more severe forms of COVID-19 and a more reliable clinical appraisal of the patient’s condition, it also resulted in a limitation on the generalizability of the findings of this study to the general COVID-19 population. The criteria for patient inclusion and exclusion allowed for better isolation of the interaction between periodontitis and COVID-19, but may also limit the generalization and extension of the results to the general population. Besides confounders such as age, sex, medications, smoking, and comorbidities, other confounders, such as BMI, socioeconomic status, and lifestyle, could not be accounted for, which may limit the validity of the results. More studies with possible standardizations on sampling and examination timings, with prospective longitudinal or controlled trial designs are required to further investigate the issues addressed in this article.