Periodontitis and Preeclampsia in Pregnancy: A Systematic Review and Meta-Analysis

Objectives A conflicting body of evidence suggests localized periodontal inflammation spreads systemically during pregnancy inducing adverse pregnancy outcomes. This systematic review and meta-analysis aim to specifically evaluate the relationship between periodontitis and preeclampsia. Methods Electronic searches were carried out in Medline, Pubmed, Embase, Lilacs, Cochrane Controlled Clinical Trial Register, CINAHL, ClinicalTrials.gov, and Google Scholar with no restrictions on the year of publication. We identified and selected observational case–control and cohort studies that analyzed the association between periodontal disease and preeclampsia. This meta-analysis was conducted following the PRISMA checklist and MOOSE checklist. Pooled odds ratios, mean difference, and 95% confidence intervals were calculated using the random effect model. Heterogeneity was tested with Cochran’s Q statistic. Results Thirty studies including six cohort- and twenty-four case–control studies were selected. Periodontitis was significantly associated with increased risk for preeclampsia (OR 3.18, 95% CI 2.26 – 4.48, p < 0.00001), especially in a subgroup analysis including cohort studies (OR 4.19, 95% CI 2.23 – 7.87, p < 0.00001). The association was even stronger in a subgroup analysis with lower-middle-income countries (OR 6.70, 95% CI 2.61 – 17.19, p < 0.0001). Conclusions Periodontitis appears as a significant risk factor for preeclampsia, which might be even more pronounced in lower-middle-income countries. Future studies to investigate if maternal amelioration of periodontitis prevents preeclampsia might be warranted.


Introduction
Preeclampsia is the onset of pregnancy-related hypertensive disorder and proteinuria arising most commonly after 20 weeks of gestation, which could lead to eclampsia and induce maternal and perinatal morbidity, and mortality. The prevalence of preeclampsia is between 2 to 8% of all pregnancies worldwide (Duley, 2009). Preeclampsia affected pregnancies had a higher risk of poor maternal outcomes including cerebrovascular bleeding, HELLP syndrome, eclampsia, poorer outcomes of their offspring including premature birth, intrauterine growth restriction, and the complications may manifest over years postpartum (Hung et al., 2018;Turbeville & Sasser, 2020).
Contributing to USD 6.4 billion short-term estimated costs for preeclamptic pregnancies in US healthcare system, USD 1.03 billion were spent on maternal healthcare and USD 1.15 billion were expended for infants born to these women while the remaining expenses were for peripartum and postpartum care (Stevens et al., 2017).
Herein, managing risk factors of preeclampsia is important to improve maternal and perinatal outcomes and lessen the burden on the health economic aspects.
Depending on geographical regions approximately 14.2 and 54.8% of pregnant women suffer from periodontal disease (Alchalabi et al., 2013;Gesase et al., 2018;Govindasamy et al., 2017). Especially, periodontitis, a more severe type of periodontal diseases affecting 11% of the pregnant women, can cause the destruction of periodontal tissue and cause systemic dissemination of bacteria and other inflammatory mediators (Bui et al., 2019;Piscoya et al., 2012). Systemic inflammatory processes triggered by focal periodontal infections have been attributed to cardiovascular, cerebrovascular diseases and respiratory diseases. (Winning & Linden, 2015) Periodontitis has independently been linked to several pregnancy complications such as preterm birth, low birth weight, and gestational diabetes (Abariga & Whitcomb, 2016;Corbella et al., 2012).
Socioeconomic status is a recognized factor associated with medical outcomes, including pregnancy outcomes (Kivimäki et al., 2020). Women with lower socioeconomic status are at a higher risk of pregnancy complications such as gestational diabetes, preterm delivery, and preeclampsia (Bo et al., 2002;Peacock et al., 1995;Silva et al., 2008). Women with high socioeconomic status have a statistically significant reduced risk of preeclampsia with an odds ratio of 0.899 (95% CI, 0.862-0.937, p < 0.001) compared to women with lower socioeconomic statuses (Ross et al., 2019). At the same time, the proportion of periodontitis in pregnancy is linked to low socioeconomic status with 42.6%, compared to high socioeconomic status with 15.0%.
Two previous meta-analyses both published in 2013 reported positive associations between preeclampsia and periodontitis with OR 2.17, 95% CI 1.38-3.41, p = 0.008 and OR of 2.79, 95% CI 2.01-3.01, p < 0.0001, but did not consider socioeconomic factors (Sgolastra et al., 2013;Wei et al., 2013). Since then further case-control and cohort studies have been published on this research topic (Lafaurie et al., 2018;Soucy-Giguère et al., 2016;Varshney & Gautam, 2014). Nonetheless, the causal relationship between periodontal disease and preeclampsia remains unclear (Kunnen et al., 2010;Lavigne & Forrest, 2020). In this review, we included all available new studies, to re-evaluate the potential association between periodontitis and preeclampsia and also to take socioeconomic factors into consideration (Australian Institute of Health and Welfare 2010. Socioeconomic variation in periodontitis among Australian adults 2004-06.)

Eligibility Criteria
The studies were screened according to the following inclusion criteria: (1) Study design was either case-control or prospective cohort study; (2) Studies analysing the association between periodontal disease and preeclampsia; (3) Study population was pregnant women without systemic diseases; (4) Preeclampsia was defined as the development of blood pressure of ≥ 140/90 mmHg after 20 weeks of gestation, combined with proteinuria of at least 1 + on midstream urine specimen or catheter specimen; (5) Periodontitis was either diagnosed ≥ 2 sites with PD ≥ 4 mm and CAL ≥ 3 mm, not on the same site or one site with PD ≥ 5 mm at the same site or evaluated the progression by clinical periodontal parameters including periodontal pocket depth, clinical attachment loss and bleeding on probing (Eke et al., 2012). The progression was measured by pocket depth (mm), clinical attachment level (mm) at baseline and delivery time; (6) Data was presented in such a way that Odds Ratio and 95% Confidence Interval could be calculated.
Studies were excluded if they did not report adequate data on periodontal or preeclamptic conditions or outcome of interest or did not meet the inclusion criteria.

Information Sources
We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines with the checklist of 27 items to conduct our study (Moher et al., 2009). We also adopted the MOOSE checklist for Metaanalysis of Observational studies (Stroup et al., 2000). A systematic search of the electronic database including Medline (from 1950), Pubmed (from 1946), Embase (from 1949), Lilacs, Cochrane Controlled Clinical Trial Register, CINAHL, ClinicalTrials.gov andGoogle Scholar (from 1990) to identify relevant articles.

Search Strategy
We used the following search terms: periodontitis, periodontal disease, preeclampsia, pre-eclampsia, pregnancy outcomes, pregnancy complications, and hypertension. The combinations of search terms were used to explore above databases. The search strategy was peer reviewed by two independent reviewers (QA and LD). The reference lists of relevant articles were also scanned for appropriate studies. No language restrictions were adopted in either the search or study selection. No search for unpublished literature was carried out. Authors were contacted for translation and information.

Study Selection
Two independent reviewers (LD and HN) reviewed the titles, abstracts and methods of retrieved results to assess for the eligibility criteria. When there was a disagreement in a selection process between reviewers, consensus with the third reviewer (QA) was obtained.

Data Extraction
Data extraction was carried out using a standardized extraction form, collecting information on the first author's name, publication year, study design, number of cases, number of controls, total sample size, country, national income group, mean age, the risk of estimates or data used to calculate the risk estimates, CIs or data used to generate CI. According to the World Bank classification of countries which is based on Gross National Income per capita, groups of national income per year are (according to World Bank classification (Country and Lending Groups & The World Bank Group, 2011)): -Low income: $995 or less; -Lower-middle-income: $996-$3945; -Upper-middle-income: $3946-$12,195; -High income: $12,196 or more.
The researchers cross-checked all extracted data and discussed if there were disagreements.

Assessment of Risk of Bias
Risk of bias was executed using the Newcastle Ottawa Scale by two reviewers (QA and LD) (Lo et al., 2014) with disagreements resolved by consensus attainment between reviewers. This scale has three components including Selection, Comparability and Outcome/Exposure assessment with maximum overall score of nine. Studies were rated as low risk of bias if they received nine score, moderate risk of bias if they received seven or eight score and high risk of bias if they received less than seven scores.

Data Synthesis
Data were imported in a statistical software (RevMan, Version 5, 2008, The Nordic Cochrane Center, The Cochrane Collaboration, Copenhagen, Denmark). Pooled Odds Ratios, mean difference, and 95% Confidence Intervals were calculated for the association between periodontitis and preeclampsia using a random effects model. The pooled effect was considered significant if p-value was less than 0.05. Forest plots for primary analysis and subgroup analysis show the raw data, Odds Ratio and CIs, Means and SDs for the chosen effect, heterogeneity statistic (I 2 ), total number of participants per group, overall Odds Ratio and Mean difference.
Subgroup analysis was carried out according to the study design (case-control or cohort), definition of periodontitis (defined by pocket depth (PD) and/or clinical attachment loss (Taghzouti et al.)), mean CAL, mean PD, national income (high-income or middle-income or low-income countries).
Heterogeneity was tested with Cochran's Q statistic, with p < 0.10 indicating heterogeneity, and quantified the degree of heterogeneity using the I 2 statistic, which represents the percentage of the total variability across studies which is due to heterogeneity. I 2 values of 25, 50 and 75% corresponded to low, moderate and high degrees of heterogeneity respectively (Higgins & Thompson, 2002). We quantified publication bias using the Egger's regression model with the effect of bias assessed using the fail-safe number method (Egger et al., 1997) The fail-safe number was the number of studies that we would need to have missed for our observed result to be nullified to statistical non-significance at the p < 0.05 level. Publication bias is generally regarded as a concern if the fail-safe number is less than 5n + 10, with n being the number of studies included in the meta-analysis (Orwin, 1983) Publication bias was assessed using Stata (16.1, Stata-Corp LLC, College Station, TX).

Risk of Bias of Included Studies
Newcastle Ottawa Scale was used to evaluate the quality of evidence of these reports. Two reviewers marked the scores for each paper based on the tool provided by the Scale. Nine studies (Ha et al., 2011(Ha et al., , 2014Jaiman et al., 2018;Khader et al., 2006;Khalighinejad et al., 2017;Lee et al., 2016;Moura da Silva et al., 2012;Pattanashetti et al., 2013;Soucy-Giguère et al., 2016) obtained the maximum score in Selection outcome while fourteen studies were marked with maximum score in the Comparability outcome and none of the studies could achieve ultimately 3 marks in the Exposure outcome. Table 2 describes the evaluation of risk of bias for this review.

Publication Bias
The funnel plot for the association between periodontitis and preeclampsia revealed the symmetry (Fig. 13). No publication bias was found.

Discussion
The aim of this meta-analysis was to re-evaluate the potential association between preeclampsia and periodontitis. The results confirm that periodontitis is a risk factor for preeclampsia, which was similar to the finding of a meta-analysis  in 2013 by Sgolastra et al. (Sgolastra et al., 2013). Our review has fifteen additional studies with three more cohort studies considerably increasing the sample size and hence generating more robust effect sizes and significance levels.
By stratifying according to study designs, periodontitis and preeclampsia showed significant associations in both case-control and cohort studies, whereas Sgolastra et al. (Sgolastra et al., 2013) could not report the statistical significance in the subgroup analysis of cohort studies (OR 2.2, 95% CI 0.66 -7.36, p = 0.2). As a review of cohort studies provides higher level of evidence compared to case-control studies (Guyatt et al., 2000) our data considerably strengthens the evidence of a positive association between preeclampsia and periodontitis. Moreover, the heterogeneity in an analysis of cohort studies in our study (I 2 = 55%, p = 0.05) was substantially lower than in a study by Sgolastra et al. (I 2 = 89%, p = 0.0001), again strengthening the reliability of our results (Sgolastra et al., 2013).
When analyzed according to the definition of periodontitis, three subgroup analysis with studies defining  periodontitis by PD alone, CAL and PD and CAL alone showed statistically significant differences, whereas the previous meta-analysis showed only significance with a subgroup analyzing periodontitis by CAL and PD. This could be explained by the number of studies included in our review was more than in the previous analysis, thus, providing a more comprehensive finding. However, according to the most recent case definition developed by the Centre for Disease Control and Prevention in partnership with the American Academy of Periodontology, the diagnostic criteria of periodontitis is at least 2 interproximal sites with the minimum of attachment loss of 3 mm and at least 2 interproximal sites with the minimum pocket depth of 4 mm (not on the same tooth) or one site with pocket depth ≥ 5 mm (Eke et al., 2012). Moreover, pregnant women who were preeclamptic had higher mean CAL and PD, however, the heterogeneity in both subgroup analysis was high, indicating significant variations among these studies in each subgroup. The high heterogeneity reported could result from the difference in the periodontal probes used in the dental examination in each study.  Several mechanisms have been proposed for the link between periodontitis and preeclampsia. Higher levels of some periodontal pathogens such as P.gingivalis and F. nucleatum were found in placenta of patients with preeclampsia (Barak et al., 2007). Moreover, inflammatory responses including the shifting of Th2 toward Th1, increasing oxidative stress, anti-angiogenic proteins, vascular endothelial growth factor receptor 1 and complement C5a could potentially enhance the development of preeclampsia (Nourollahpour Shiadeh et al., 2017). Ananth et al. has reported the association between intrauterine growth restriction and maternal periodontitis (Ananth et al., 2018). Since severe and early onset preeclampsia were associated significantly with fetal growth restriction, this could contribute to the mechanism underlying the association between preeclampsia and periodontitis (Odegård et al.,  . Furthermore, the mechanisms might be a reflection of dietary patterns. Recently, some evidence has indicated that pathogenesis of preeclampsia involves maternal gut microbiota, specifically, high-fiber diet which promote short chain fatty acid production and are associated with reduced risk of preeclampsia (Hu et al., 2019). Similarly, high-fiber foods such as fruit and grains have been linked to the reduction of the progression of periodontal disease, suggesting the role of dietary intake in the potential relationship between preeclampsia and periodontal disease (Hu et al., 2019;Schwartz et al., 2012). However, future studies are required to elucidate these hypotheses.
When analyzing the association between periodontitis and preeclampsia according to national income, this review revealed the significant difference in the subgroup analysis of high-income and upper-middle-income countries (OR = 2.67 and OR = 2.40, respectively). Moreover, the subgroup analysis with lower-middle-income countries, which generated an Odds ratio of 6.70, indicated the considerable significance in the relationship between periodontitis and preeclampsia in this specific country group. The high heterogeneity was observed in the subgroup of lower-middle-income countries which implies the variation between included countries. Countries were categorized as lower-middle-income using national gross income per capita, thereby being subject to variation even within each individual country. In other words, using national income as a proxy may result in this variation and therefore, suggested the individual level approach for future studies to tackle this issue. Moreover, lower middle-income countries have poorer oral health condition compared to upper-middle and high-income country groups, which may indicate the inequalities in oral health care (Bastani et al., 2021;Watt & Sheiham, 1999). Inequalities can stem from unjust provision of services or inappropriate access and become more pronounced by the fact that most dental treatment is funded by out-of-pocket payments (Listl et al., 2015).
Socioeconomic inequalities in access to oral health care accounted for 60% in lower-middle-income countries (Hosseinpoor et al., 2012). Therefore, improving access to oral health services for pregnant women in lower-middle-income countries is notably important to lessen the risk of having preeclampsia. Additionally, the allocation of resources for oral care might need more investigations and strategic management to eliminate the disparities in oral health care (Arevalo & Tomar 2019). Social and cultural determinants include biological, behavioral, cultural, social and political aspects should also be focused to thoroughly eradicate the inequalities (Patrick et al., 2006).
We used Newcastle Ottawa Scale to evaluate the risk of bias and found twenty studies with moderate risk of bias and ten remaining studies with a high risk of bias. Exposure bias in this study was due to the non-response rate which was not described clearly in these studies. No publication bias was detected. Sensitivity analysis resulted in no change to the finding of the study.  The strength of this systematic review and meta-analysis includes the large sample size of 9650 subjects. Six cohort studies comprising 2840 subjects were analyzed and revealed the statistically significant difference. Because systematic reviews of prospective cohort studies generate more reliable evidence, our study provided an updated systematic review and meta-analysis and confirmed the association between periodontitis and preeclampsia (Hillier et al., 2011). Furthermore, by stratifying into subgroup analysis of national income, our review has pointed out the association between these two diseases differed according to economical inequalities, thus, providing recommendation for health policy improvement. Pregnant women in low socioeconomical areas should be given access to oral healthcare services and encouraged to have their periodontal health checked and treated during pregnancy to potentially lower the risk of preeclampsia and other pregnancy complications. Jeffcoat et al. reported non-surgical periodontal therapy could significantly reduce the medical costs for pregnant women by 73.7% (Jeffcoat et al., 2014). We acknowledged few limitations in our study. Firstly, the heterogeneity of the overall analysis for the association between periodontitis and preeclampsia was high, pointing out the variations among studies included. The synthesis of cohort and case-control studies in our review may explain for this high heterogeneity. Secondly, the general consensus in the definition and diagnosis of periodontitis was not clear enough which could influence the results of our meta-analysis. Additionally, because the intraoral examinations were conducted at different time points, the diagnosis of periodontitis may be impacted. Deteriorated periodontium was observed more in the third trimester compared to the second trimester and overall periodontal health was improved postpartum (González-Jaranay et al., 2017;Kashetty et al., 2018). Thereby, future clinical studies should consider the time point of seven days of the elivery when conducting periodontal examinationsand confirm our results. Moreover, because our analysis was based on national income, new research with individual-level data of socioeconomic factors is recommended for a more informative conclusion.

Conclusions and Implications
This meta-analysis confirms previous findings of an association between periodontitis and preeclampsia. However, this study includes fifteen more recent publications, which resulted in a larger effect size of the association, specifically, for lower-middle-income countries in comparison to high and upper-middle-income countries. Our results warrant future studies to investigate the mechanisms of this association and whether targeted interventions to prevent or treat periodontitis preconception or during pregnancy can lead to better pregnancy outcomes.
Author Contributions QAL Data curation, data analysis, quality assessment of articles, writing-original, and draft preparation. RA methodology, writing-editing, and supervision. KMC methodology, writing-editing, and supervision. NTN data curation, methodology, and writing-review. LDTY data curation, and quality assessment of articles. HNV data curation, and quality assessment of articles. AY writingreview and editing. GC writing-review and editing. JE methodology, writing-review and editing, and supervision. RN conceptualization and project administration.
Funding Open Access funding enabled and organized by CAUL and its Member Institutions. This research was funded by the Scholarship provided by the Chair of Lifespan Oral Health to Quynh Anh Le.
Data Availability Data will be available upon request.
Code Availability Not applicable.

Conflict of interest
The authors report no conflict of interest.
Ethical Approval Not applicable.

Consent for Publication Not applicable.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.