Each year, about 15 million infants worldwide are born preterm (before 37 weeks of gestation), and these preterm babies typically have low birth weight (LBW, <2,500 g) (1). Preterm birth (PTB) is the leading cause of neonatal mortality, morbidity, and developmental loss (2). Advances in obstetric care have not altered the rates of PTB, and it is estimated that 9.6% of worldwide births are preterm (3). The highest rates of PTB are in Africa (11.9%) and North America (10.6%), and the lowest rates are in Europe (6.2%) (3). However, the underlying causes of PTB are still not entirely clear, thus an accurate identification of risk factors for PTB that are amenable to intervention would have far-reaching and long-lasting impact.

Of the multiple risk factors for PTB, maternal infection is identified consistently. Periodontal disease is a highly prevalent infectious and inflammatory disease of tooth-supporting tissues and if untreated can lead to oral disabilities (Figure 1A) (4). Periodontal disease is caused mainly by gram-negative microaerophilic and anaerobic bacteria that colonize the subgingival area and produce significant amounts of proinflammatory mediators (5). Periodontal disease includes gingivitis and periodontitis. Gingivitis is the presence of gingival inflammation without loss of connective tissue attachment. Periodontitis is the presence of gingival inflammation at sites where there has been apical migration of the epithelial attachment onto the root surfaces accompanied by the loss of connective tissue and alveolar bone (Figure 1B) (6). In the last two decades, many studies have examined the relationship between periodontitis and PTB. Periodontitis may be a risk factor for PTB due to the presence in the bloodstream of bacteria and proinflammatory cytokines during infection that can affect distant organs (7). However, epidemiological studies and intervention trials have reached contradictory conclusions about the relationship between periodontitis and PTB. The aim of this review is to summarize the current evidence from epidemiological, animal, and in vitro studies, as well as intervention trials and to propose new ideas about the link between periodontitis and PTB.

About one-third of all PTB are caused by preterm labor (uterine contraction) and one-third are due to the premature rupture of membranes (PROM); the remaining cases are due to other pregnancy complications such as induced labor (of which preeclampsia is the major indication) (8). In the last two decades, numerous epidemiological studies have examined the link between periodontitis and PTB, including cross-sectional, case–control, and cohort studies.

In a cross-sectional study, or census, data are collected at a defined time and is used to assess the prevalence of chronic or acute conditions, the results of intervention, or the causes of disease. In the last 5 years, several cross-sectional studies (9–11) reported a correlation between periodontitis and PTB/LBW. In a study published in 2016 (9), women with PTB were found to have worse periodontal parameters and significantly increased gingival crevicular fluid (GCF) levels of IL-6 and prostaglandin E₂ (PGE₂) compared with women who experienced full-term birth. Based on significant correlations between serum PGE₂ level and probing depth, clinical attachment loss (CAL, Figure 1C), and GCF TNF-α in PTB, periodontitis may increase the risk of labor triggers and hence contribute to preterm labor onset. However, in 2016 Martinez-Martinez et al. (12) suggested that PTB is a multifactorial condition and that periodontitis and the presence of periodontal pathogens are not sufficient to trigger PTB.

In case–control studies, mothers with PTB are identified and their periodontitis history is determined and compared with that of healthy control subjects. Of the 14 case–control studies published in the last 5 years, 12 (13–24) reported an association between periodontitis and PTB, LBW, or preterm LBW (PLBW), and 2 (25, 26) found no association. In a study published in 2015 (23), mothers in the periodontitis group with single delivery had an eightfold higher chance of delivering a LBW infant compared to those in the control group. The mothers in the periodontitis group with multiple deliveries delivered PTB infants with an eightfold higher frequency and LBW infants at a 10-fold higher frequency compared to the mothers in the control group.

In cohort studies, investigators monitor pregnant women to determine if those with periodontitis demonstrate a higher incidence of PTB than those without periodontitis. Of the 11 published cohort studies in the last 5 years, 7 (27–33) reported an association between periodontitis and PTB, LBW, or PLBW, and 4 (34–37) revealed no association. A hospital-based prospective study published in 2016 (33) comprising 790 pregnant women found that periodontitis was a risk factor for PTB and an independent risk factor for LBW. Recently periodontitis was also found to be associated with preeclampsia (38) and PROM (39, 40), common causes of PTB.

In the last 5 years, most literature and systematic reviews reported an association between periodontitis and PTB (41–46). A meta-analysis published in 2016 (41) assessed case–control studies reporting periodontal status and pregnancy outcomes. The computed risk ratio for periodontitis was 1.61 for PTB using data from 16 studies, the risk for LBW was 1.65 based on 10 studies, and the risk for PLBW was 3.44 based on 4 studies. A systematic and evidence-based review in 2012 (46) focused on the association of periodontitis and PTB and LBW and found 62 relevant studies that suggested that periodontitis may be a potential risk factor for PTB and LBW.

Different epidemiological data may have reached different conclusions due to the following reasons: (1) many studies used different definitions of periodontitis and adverse pregnancy outcomes, for instance the use of PLBW as a composite outcome, or PTB versus LBW, terms that reflect different disease severities and pathologic entities. (2) The risk factors of PTB may be similar to the risk factors for periodontitis (ethnicity, tobacco use, and socioeconomic and educational levels) and thus may confound the association between periodontitis and PTB. In a prospective case–control study (47), PTB was associated with periodontitis when the USA (48), but not the European (49), definitions were used. Therefore, future studies should employ both continuous and categorical assessment of periodontal status and control for confounding factors. Additionally, the further use of the composite outcome PLBW is not encouraged.

Periodontitis is one of the most common chronic infectious diseases and is caused mainly by gram-negative microaerophilic and anaerobic bacteria that colonize the subgingival area and produce significant amounts of proinflammatory mediators, mainly IL-1β, IL-6, PGE₂, and TNF-α (4). Periodontitis may act as a distant reservoir of both microbes and inflammatory mediators that may influence pregnancy and contribute to induction of PTB (Figure 2). These two potential mechanisms for how periodontitis can affect PTB are described more fully below.

The possible roles of periodontitis in PTB have also been explored using experimental animal models. In separate studies, periodontal bacteria were injected into a small chamber in pregnant animals, allowing the establishment of a site of infection distant to the fetal–placental unit to mimic a periodontal infection in a reproducible and simplified manner. These results revealed that maternal infections with periodontal pathogens increase pregnancy complications.

Dental infection of mice with P. gingivalis significantly increased levels of circulating IL-1β, IL-6, IL-17, and TNF-α (74). Defects in the placental tissues of P. gingivalis-infected mice included degenerative changes in endothelial and trophoblast cells, increased placental detachment, and PROM, and P. gingivalis was detected in placental tissues by PCR and immunohistochemistry (74). The P. gingivalis-infected group delivered at gestational day (gd) 18.25 versus gd 20.45 for the non-infected control group (p < 0.01), with pups exhibiting LBW compared to controls (p < 0.01) (74). In another study (75), mice with periodontitis induced by using an inoculum of P. gingivalis and F. nucleatum exhibited increased circulating levels of IL-6 and IL-8. Similarly, Miyoshi et al. (76) found high levels of contractile-associated proteins and ion channels in the myometrium of PTB model mice with chronic odontogenic P. gingivalis infection. In murine models, F. nucleatum translocated and caused intrauterine infections (77) and Campylobacter rectus significantly decreased fetoplacental weight (78). In a baboon model, a significantly greater frequency of the periodontitis group neonates had decreased gestational age and LBW (79). Spontaneous abortion/stillbirth/fetal demise was increased in the periodontitis group versus the control group (79). Also, combined oral infection of mice with P. gingivalis and C. rectus significantly reduced overall fecundity compared to controls (80). Overall, most animal studies reported a harmful impact of periodontitis on pregnancy outcome. However, a study performed by Fogacci et al. (81) found no increased risk for PTB or LBW in Wistar rats with induced periodontitis.

In addition to the data from animal studies, in vitro experiments have been conducted to explore the molecular mechanisms underlying periodontitis-induced PTB. In most in vitro models, periodontal pathogens/byproducts were used to infect placental tissues or trophoblast cells/cell lines. These experiments were designed to mimic an in vitro periodontal infection in a simplified and reproducible manner to allow investigation of the interaction between periodontal pathogens and placental tissues/cells.

Riewe et al. (82) investigated the transcriptional responses after infection with P. gingivalis in extravillous trophoblast (HTR8) cells derived from the human placenta, and found that over 2,000 genes were differentially regulated. P. gingivalis induced IL-8 and IFN-γ secretion (83), apoptosis, and arrest in the G(1) phase of the cell cycle (84, 85) in HTR8 cells. Increased IFN-γ secretion and Fas expression occurred in P. gingivalis-induced apoptosis of HTR8 cells via the ERK1/2 pathway (86). In normal human term fetal membrane explants, P. gingivalis may significantly increase TLR7 expression (87).

Porphyromonas gingivalis has bioactive components on the cell surface, including lipopolysaccharide, capsules, and fimbriae. P. gingivalis lipopolysaccharide induces the production of IL-6 and IL-8 via TLR2 in chorion-derived cells (50) and increases expression levels of IL-8, TNF-α, and COX-2 in HTR8 cells in an NF-κB-dependent manner (74). Interestingly, Komine-Aizawa et al. (88) reported that although there is limited direct pathogenic effect of P. gingivalis lipopolysaccharide on trophoblast invasion, concurrent smoking reduces trophoblast invasion into the myometrium and thus inhibits maternal vascular reconstruction.

In human placental trophoblast-like BeWo cells, the presence of A. actinomycetemcomitans lipopolysaccharide led to increased levels of cytochrome c, caspase-2, caspase-3, caspase-9, and BCL2-antagonist/killer 1 mRNA and decreased levels of B-cell CLL/lymphoma 2, BCL2-like 1, and catalase mRNA, consistent with the activation of the mitochondria-dependent apoptotic pathway (89). Additionally, C. rectus was reported to effectively invade BeWo cells and upregulate both mRNA and protein levels of IL-6 and TNF-α in a dose-dependent manner (78). Therefore, there is significant evidence that periodontal pathogens and byproducts can induce inflammation and/or apoptosis in placental tissues and cells.

Periodontal treatment usually refers to non-surgical periodontal therapy that can improve periodontal health and is defined as plaque removal, plaque control, supragingival and subgingival scaling, root surface debridement, and the adjunctive use of chemical agents. To evaluate periodontitis as a risk factor for PTB, intervention studies were conducted to evaluate the effect of periodontal treatment on the risk of PTB. In these studies, women with preexisting periodontitis were randomly divided into two groups. One received periodontal treatment during or before pregnancy, and the other did not. In this type of study, the researchers could assess if periodontitis was an independent risk factor for PTB by determining if periodontal treatment decreased the incidence of PTB.

In a study published in 2015 (90), the mean gestational age in the periodontal treatment group (treatment performed during the second trimester of the gestational period) was 35.57 ± 2.40 versus 34.17 ± 2.92 weeks in the non-treated group (p < 0.05). The treatment group showed a statistically significant reduction in mean CRP levels after delivery compared to baseline values; the control group showed no significant reduction in CRP levels. Another study (91) suggested that periodontal treatment during pregnancy is not only safe for both the child and the mother, but also provides beneficial effects for pregnancy and embryo-fetal development, leading to reduced morbidity and mortality in PTB infants. Macedo et al. (92) also reported an association between a low number of daily tooth brushings and PTB. However, in other recent studies (93–96), periodontal treatment performed on pregnant women was not found to be efficacious in reducing PTB or LBW.

Data from recent systematic reviews are also contradictory. Several reviews (97–99) reported that the risks of perinatal outcomes could be potentially reduced by periodontal treatment in pregnant women, but other reviews (100–103) suggested that periodontal treatment during pregnancy was not an efficient way to reduce the incidence of PTB. However, the evidence was not conclusive due to confounding effects and risks of random errors and bias. Thus, further randomized clinical trials are still necessary.

The preventive effectiveness of periodontal treatment for PTB has still not been established, because it is influenced by many factors such as the diagnostic criteria of periodontitis, microbial community composition, severity of disease, treatment strategy, treatment efficiency, and the timing of treatment during pregnancy (the pre-pregnancy period or during the first or second trimester). Jeffcoat et al. (104) confirmed that decreased PTB may depend on the success of periodontal treatment. In this study of 322 pregnant women with periodontitis, 162 were randomly assigned to receive only oral hygiene instruction and served as the untreated control group, whereas the remaining 160 received scaling and root planing treatment as well as oral hygiene instruction. No significant difference was found between the incidence of PTB in the periodontal treatment group and the control group. However, a logistic regression analysis showed a significant and strong relationship between successful periodontal treatment and full-term birth. Subjects refractory to periodontal treatment were significantly more likely to have PTB. Similarly, in another study (105), periodontal treatment during pregnancy reduced the levels of IL-1β, IL-6, IL-10, and IL-12 in GCF and improved dental parameters. Additionally, the severity of periodontitis was significantly associated with an increased risk of babies born small for gestational age, but no changes in pregnancy-related outcomes were observed following periodontal treatment. These studies suggest the need for the next randomized controlled trials to standardize methodological criteria and utilize a more precise definition of periodontitis. Additionally, for better statistical power, studies should preferably be performed as multicenter studies, and include a large number of participants. Finally, the resulting success or failure of periodontal treatment must be considered.

Preterm birth is the leading cause of infant morbidity and mortality, but classical risk factors explain only one-third of PTB cases, and current intervention strategies have not led to an appreciable reduction of PTB. Therefore, it is necessary to explore mechanisms of causality and generate new hypotheses using an integrated approach. This should be done with increased collaboration among research groups, and using more comprehensive theoretical–methodological approaches to formulate more effective intervention strategies and to detect new risk factors. Although intervention during pregnancy has not consistently been correlated with a reduction in PTB rates, oral health maintenance is an important part of preventive care that is both effective and safe throughout pregnancy and should be supported before and during pregnancy.

Here, for the first time, we describe four possible models of periodontitis in PTB: (1) periodontitis is an independent risk factor for PTB; (2) severe and/or generalized periodontitis promotes PTB; (3) periodontitis only promotes PTB for pregnant women who are young or HIV-infected or have preeclampsia, pre-pregnancy obesity, or susceptible genotypes; and (4) periodontitis has no significant effect on PTB. Because contradictory epidemiological data continue to emerge (model 1 versus 4), future studies should try to test the second and third models, which may, to some extent, explain the conflicting epidemiological data. Although intervention during pregnancy has not consistently been correlated with a reduction in PTB rates, oral health maintenance is an important part of preventive care that is both effective and safe throughout pregnancy and should be supported before and during pregnancy.

HR contributed to the literature search, interpretation, writing, and proofreading of the manuscript. MD designed the study and made the ultimate decision on the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.