Pregnancy is a unique period during a woman’s life. Throughout this period many changes take place in the mother’s body reflecting her adaptation to the new life that is growing inside the womb. However, in some cases pregnancy does not progress uneventfully and pregnancy complications may develop. Adverse pregnancy outcomes (APOs) are the leading causes of maternal and foetal morbidity and mortality. Some of the most common APOs include preterm birth (PTB), low birth weight (LBW) and pre-eclampsia.
Preterm birth (PTB) is defined as a live birth of an infant before 37 weeks of gestation. PTB infants can be further subdivided in extremely preterm (born before 28 weeks), severely premature (between 28 and 31 weeks), moderately premature (between 32 and 33 weeks) and near term (between 34 and 36 weeks). PTB is a serious pregnancy complication since it has been estimated to be the major cause of 28% of neonatal deaths worldwide1. The rate of PTB varies considerably among populations. In the USA one in every ten live births is preterm2, whilst in non-Hispanic black infants this percentage reaches almost 17%3. In Europe the average prevalence of PTB is 5% to 9%3. PTB may be medically indicated, or spontaneous, or occur after premature rupture of the amniotic membranes. Here, intrauterine inflammatory mediators have been implicated in the initiation of this process.
LBW is defined as weight at birth of less than 2.5 kg (5 lb 8 oz), approximately one-third less than the average weight of a term infant (3.5 kg/7 lb), while very low birth weight (VLBW) is defined as birth weight of less than 1.5 kg. The main reasons for LBW or VLBW are incomplete foetal development during pregnancy and preterm labour. Therefore, a distinction should also be made between intrauterine growth restriction and size at gestational age, which is defined as weight of less than the 10th percentile at any given gestational age.
Pre-eclampsia is a maternal multi-organ disease of unknown aetiology that appears clinically after the 20th week of gestation. Usually, it manifests in the pregnant woman as hypertension (higher than 140/90 mmHg) and proteinuria, while the foetus may develop intrauterine growth restriction. If left untreated, pre-eclampsia may result in an increased risk of maternal and perinatal mortality (PNM)3. It is a disorder unique to pregnancy, with a prevalence of approximately 2% to 7% in developed countries, while in developing countries, its prevalence exceeds 10%. Pre-eclampsia risk factors include advanced maternal age, multi-foetal pregnancies, obesity, hypertension, renal disorders and diabetes mellitus4. However, mounting evidence suggests that infection may also be critical in the pathogenesis of pre-eclampsia, both in terms of its initiation and its potentiation.
Despite the advances in health care, APOs still remain an important public health problem. For instance, besides maternal and infant mortality, surviving PTB infants often face multiple lifelong challenges, including respiratory distress, impaired motor skills, cognitive and intellectual impairment, learning difficulties and cardiovascular and metabolic disorders5. Understanding the financial and social impact that these conditions may induce throughout life, it is reasonable to suggest that pregnancy complications have significant consequences not only in terms of the health of the affected babies and mothers, but also for their families and society in general6. Therefore, early identification of risk factors during prenatal care is important for prevention and control.
Since an increase in the local and systemic inflammatory burden has, to a certain extent, been implicated in the aetiopathogenesis of these APOs, much focus has been placed on the identification of sources that could exacerbate these processes. Vaginal infections have been the centre of these investigations, although more distant infections have also drawn the attention of researchers. The periodontal diseases gingivitis and periodontitis are such distant infections. Besides the local inflammatory response that is launched against the oral microbiome, a systemic inflammatory response is also initiated during periodontal diseases. Moreover, the entrance of periodontal bacteria to the blood circulation through the inflamed tissues and therefore the possibility of their dissemination to different sites of the body brings this distant infection closer to the foetal-placental unit.
Therefore, over the last 20 years many epidemiological studies have taken place in order to assess whether periodontal diseases are associated with common APOs. In addition, intervention studies and especially randomised clinical trials (RCTs) have investigated whether periodontal therapy during pregnancy can reduce the risk of APOs. Thus, they assessed whether this possible association is causal in nature. At the same time, many mechanistic studies in humans and animal models have tried to elucidate the possible biological pathways that may connect periodontal diseases with the aetiopathogenesis of APOs.
In the general context of a possible association between periodontal disease and pregnancy complications several investigators have also sought to evaluate whether periodontal diseases may affect fertility. The estimated prevalence of infertility among couples of reproductive age is 15%, and at least 40% to 50% of the cases are associated with male factor infertility7. However, in about half of the cases, the aetiology remains unexplained. The aetiology of infertility is considered multifactorial, and several risk factors have been associated with this condition. The inability of women to ovulate, structural problems in the fallopian tube or uterus, GU infections in both females and males and bacteriospermia are major predictors for infertility. To date, knowledge of a possible association between periodontal diseases and fertility is based upon a limited number of studies.
Finally, during pregnancy significant fluctuations in the levels of sex hormones, such as oestrogens and progestogens, occur. Since periodontal tissues possess receptors for these hormones it is likely that they have an effect on periodontal status. Indeed, an increase in gingival inflammation during pregnancy has been observed and may be the result of hormonal changes, but also of alterations of oral biofilm composition, and increased vascularity.
7.2.1 Findings from epidemiological studies
Since the first epidemiological report by Offenbacher and co-workers in 19968, many others have followed, attempting to elucidate whether periodontal diseases are associated with various pregnancy complications. So far, more than 100 such studies have been published, 40 of which were in the last 5 years. From all the existing APOs the focus has been placed on PTB, LBW, PTB and LBW (PTLBW) and pre-eclampsia, whilst a recent study has investigated the frequency of periodontitis amongst women suffering from preterm prelabour rupture of membranes (PPROM)9. Numerous reviews have evaluated the results of these epidemiological studies. Indeed, in 2013 a thorough systematic review exploring the available evidence was published as part of the proceedings of a workshop on periodontitis and systemic diseases that was held jointly by the European Federation of Periodontology (EFP) and the American Academy of Periodontology (AAP)10. Four years later, the EFP conducted an update of the existing knowledge that was prepared by Petrini et al11. These investigators divided the studies based on their design, into case-control, cross-sectional and prospective studies, and performed a systematic review and meta-analysis. Since meta-analyses are considered to provide the highest level of evidence to demonstrate a possible association between two conditions, the results of this study on the association of periodontal disease with PTB, LBW, PTLBW and preeclampsia are summarised in Table 7-1.
Table 7-1 In vivo human studies investigating the association of periodontal disease and APOs*
|Overall outcome (summary)
|Significant associations were found between APOs (pre-eclampsia, LBW, PTB) and maternal periodontitis.
|Bassani et al12
|Turton and Africa13, Gesase et al14
|Vogt et al15, Srinivas et al16, Al Habashneh et al17, Kumar et al18, Santa Cruz et al19, Tellapragada et al20, Lohana et al21
|No association of periodontitis with PPROM and associated microbial invasion of the amniotic cavity and intra-amniotic inflammation.
|Radochova et al22
|Radochova et al9
|The evidence on PTB suggests a higher tendency for periodontitis in women who deliver preterm. Prospective studies confirm these finding with a high relative risk. However, the data are weakened by the extreme heterogeneity. Meta-analysis data*: Case-control: OR 1.38 (95% CI 0.95–1.82, P < 0.01). Cohort: pooled adjusted RR 1.93 (95% CI 1.12–2.73, P < 0.01).
|Moore et al23, Wood et al24, Vettore et al25,26, Nabet et al27, Iwanaga et al28, Jarjoura et al29, Bosnjak et al30, Radnai et al31, Le et al32, Siqueira et al33, Giannella et al34, Piscoya et al35, Martínez de Tejada et al36, Abati et al37, Ye et al38, Pattanashetti et al39, Macedo et al40, Bulut et al41, Jain et al42, Govindaraju et al43, Márquez-Corona et al44
|Lunardelli and Peres45, Heimonen et al46, Santos-Pereira et al47, Toygar et al48, Guimarães et al49, Kothiwale et al50, Muwazi et al51, Martínez-Martínez et al52, Perunovic et al53, Kopycka-Kedzierawski et al54
|Offenbacher et al55, Agueda et al56, Rakoto-Alson et al57, Moore et al58, Wang et al59, Hassan et al60, Kruse et al61, Dörtbudak et al62
|Overall evidence on LBW supports a possible association with periodontal disease. However, data from prospective studies are not statistically significant and those from cross-sectional studies are highly heterogenous. Meta-analysis data*: Case-control: pooled OR 1.03 (95% CI 0.27–1.79, P < 0.01). Cross-sectional: pooled OR 2.36 (95% CI 1.18–3.54), (P < 0.01). Cohort: pooled RR 2.31 (95% CI 2.32–4.97, P = 0.21).
|Moore et al23, Cruz et al63, Vettore et al25,26, Martinez de Tejada et al36, Abati et al37, Haerian-Ardakani et al64, Mathew et al65, Jacob and Nath66, Gomes-Filho et al67, Souza et al68, Martins Moliterno et al69
|Lunardelli and Peres45, Toygar et al48, Siqueira et al33, Guimarães et al70, Kothiwale et al50, Muwazi et al51
|Marin et al71, Farrell et al72, Agueda et al56, Saddki et al73, Rakoto-Alson et al57, Wang et al59
|Overall evidence on PTLBW is inconclusive due to the poor quality of data. Meta-analysis data*: Meta-analysis was not performed because of the absence of adjusted data.
|Offenbacher et al8, Gomes-Filho et al74, Toygar et al48, Davenport et al75, Buduneli et al76, Noack et al77, Vettore et al25,26, Khader et al78, Mesa et al79,80, Blanc et al81, Kayar et al82, Fogacci et al83, Reza Karimi et al84, Kothiwale and Panwar85, Krüger et al86
|Agueda et al56, Ercan et al87
|Maternal periodontitis is significantly associated with pre-eclampsia, despite the significant heterogeneity among the data. Meta-analysis data*: Case-control: pooled OR 2.85 (95% CI 0.57–5.12, P < 0.01). Cohort: Meta-analysis was not performed because of the absence of adjusted data.
|Canakci et al88, Contreras et al89, Cota et al90, Siqueira et al91, Nabet et al27, Taghzouti et al92, Khader et al93, Abati et al37, Moura da Silva et al94, Chaparro et al95, Pattanashetti et al39, Sayar et al96
|Hirano et al97
|Riché et al98, Boggess et al99, Horton et al100, Kumar et al101, Ha et al102
*Data from one of the most recent meta-analyses (Petrini et al11).
APO = adverse pregnancy outcome; LBW = low birth weight; OR = odds ratio; PPROM = preterm prelabour rupture of membranes; PTB = preterm birth; PTLBW = preterm low birth weight; RR = relative risk.
So far, 35 studies have investigated the possible association between periodontal disease (periodontitis or gingivitis) and LBW. From these studies, 24 show a statistically significant association, while the remaining studies did not report significant results. PTB and periodontitis were assessed in a total of 47 studies of which 25 demonstrated a statistically significant correlation. PTLBW was significantly related with periodontitis in 7 out of 18 studies. Finally, 13 out of 18 studies found a statistically significant association between maternal periodontitis and pre-eclampsia.
A more in-depth analysis of these studies by one of the most recent meta-analyses11 as well as by a 2018 umbrella review of meta-analyses103 revealed that the overall evidence on LBW supports a possible association with periodontitis. However, data from prospective studies are not statistically significant and those from cross-sectional studies are highly heterogenous due to methodological inconsistencies among the studies, as will be described later. The evidence on PTB suggests a higher tendency for periodontitis in women who deliver preterm. Prospective studies confirm these findings, reporting a high relative risk (RR) of PTB in patients with periodontitis. However, the data again are weakened by the marked diversity. Maternal periodontitis is significantly associated with pre-eclampsia, despite the heterogeneity among the data. Therefore, studies on APOs and periodontitis indicate that there might be some association. However, the strength of this association is modest and hampered by the significant heterogeneity. These conclusions are consistent with those of previous reviews including the proceedings of the EFP/AAP workshop on periodontitis and systemic diseases10,11,104.
7.2.2 Findings from intervention studies
Due to their design, epidemiological studies are considered association studies since they only reveal that two conditions co-exist and therefore are associated together. However, whether this association, if present, is causative in nature, i.e. whether periodontal disease contributes to pregnancy complications, can be evaluated through intervention studies. In these investigations causality can be implied if elimination of the exposure (periodontal disease) reduces the risk of APOs.
To date, 16 independent RCTs have been published. Table 7-2 summarises the effects of non-surgical periodontal intervention during pregnancy on APOs. From this table it is obvious that not all studies evaluated the same APOs. PTB was evaluated as the main outcome in all studies, LBW was assessed in 12 studies, while PTLBW was determined in 7 RCTs. Other pregnancy complications such as pre-eclampsia, neonatal intensive care admissions, neonatal deaths, and Apgar scores (evaluation of a newborn baby on the five criteria appearance, pulse, grimace, activity, respiration) were reported as secondary outcomes in three of these studies108,111,112.
Table 7-2 Main effects of non-surgical periodontal intervention during pregnancy on APOs
|Number of patients randomised
|Effect of intervention in at least one APO
|Lopez et al105
|Significantly higher incidence of PTB and PTLBW in the non-treated control group.
|Jeffcoat et al106
|No significant differences in PTB (< 37 wk or < 35 wk) among the treated and non-treated control group. Similar results when metronidazole was combined with SRP.
|Lopez et al107
|Significantly higher incidence of PTB and PTLBW in the non-treated control group.
|Michalowicz et al108
|No significant differences in PTB, birth weight, rate of small for gestational age and preeclampsia among the treated and non-treated control group.
|Offenbacher et al55
|Significantly higher incidence of PTB in the non-treated control group.
|Sadatmansouri et al109
|Significantly higher incidence of PTLBW in the non-treated control group. No significant differences in PTB and LBW among the treated and non-treated control group.
|Tarannum and Faiduzzin110
|Significantly higher incidence of PTB and LBW in the non-treated control group.
|Newnham et al111, 2009
|No significant differences in PTB, birth weight and pre-eclampsia among the treated and non-treated control group.
|Offenbacher et al112
|No significant differences in PTB (< 37 wk, < 35 wk and 32 wk), birth weight and pre-eclampsia among the treated and non-treated control group.
|Radnai et al113
|Significantly higher incidence of PTB, LBW and PTLBW in the non-treated control group.
|Macones et al114
|No significant differences in PTB (< 37 wk, < 35 wk and indicated PTB), and LBW among the treated and non-treated control group.
|Oliveira et al115
|No significant differences in PTB, LBW and PTLBW among the treated and non-treated control group.
|Pirie et al116
|No significant differences in PTB and among the treated and non-treated control group.
|Weidlich et al117
|No significant differences in PTB, LBW and PTLBW among the treated and non-treated control group.
|Reddy et al118
|No significant differences in PTB and LBW among the treated and non-treated control group.
|Caneiro-Queija et al119
|No significant differences in PTB and/or LBW.
LBW = low birth weight; PTB = preterm birth; PTLBW = preterm low birth weight; SRP = scaling and root planing.
Although there is a unanimous agreement in all studies regarding the safety of non-surgical periodontal interventions during the second trimester of pregnancy, the results from the RCTs are controversial. However, the majority of studies (nine out of 15) demonstrate that periodontal intervention has no effect on the risk of developing any of the APOs. Specifically, five55,105,107,110,113 out of 15 studies showed a positive effect of periodontal treatment on reducing PTB, while only two110,113 out of nine revealed a reduction in LBW in the treatment group. All other studies assessing birth weights reported no differences among the treatment and the control groups, although LBW incidence was not reported. The three studies reporting data for pre-eclampsia did not find any statistical differences among the treatment and the control groups108,111,112.
Since the results from the RCTs are not consistent, several systematic reviews and meta-analyses have attempted to evaluate the quality of these studies and assess whether a safe conclusion can be extracted. Therefore, they performed risk of bias assessments for the individual RCTs in order to distinguish the high-quality studies from those of lower quality. Furthermore, most of the meta-analyses performed subgroup analyses to determine whether specific pregnancy populations may benefit or not from periodontal treatment. In summary, very few meta-analyses demonstrated a positive effect of periodontal treatment in reducing the incidence of PTB or LBW120–122, whereas others found weak, unclear or no evidence123,124. Indeed, when only high-quality studies were analysed, none of these studies revealed a benefit of periodontal therapy in decreasing the risk of PTB or LBW. However, when subgroup analysis included only pregnant women at high risk for pregnancy complications, the majority of meta-analyses showed that periodontal treatment reduced the risk for both these APOs120,125,126. At the same time, none of the very few meta-analyses demonstrated overall, or in subgroups analysis, a significant effect of periodontal therapy during pregnancy on PNM110,126,127.
Further insight regarding the effects of treating periodontal disease in pregnant women in order to prevent or reduce perinatal and maternal morbidity and mortality can be obtained from a recent (2017) and high-quality Cochrane Database Systematic Review128. According to the authors, all of the 15 RCTs (7161 participants) included in the systematic review were at high risk of bias mostly due to lack of blinding and imbalance in baseline characteristics of participants. The meta-analysis showed no clear difference in PTB between periodontal treatment and no treatment. There was also low-quality evidence that periodontal treatment may reduce LBW. It was unclear whether periodontal treatment leads to a difference in PNM and pre-eclampsia and there was no evidence of a difference in ‘small for gestational age’ when periodontal treatment was compared with no treatment. Finally, maternal mortality and adverse effects of the intervention did not occur in any of the studies that reported on either of the outcomes.
Taken together it is apparent that there is no evidence from these studies to support that non-surgical periodontal therapy during pregnancy may alter the incidence of PTB, LBW and PNM. However, there is limited evidence of a positive effect of periodontal treatment in decreasing PTB and LBW rates in women at high risk of APOs.
7.2.3 Inconsistency among epidemiological and intervention studies
A striking observation from the various epidemiological studies is the significant discrepancy in their results. Although more than 100 studies have been published and although a large number of these studies show a positive association between periodontal disease and APOs, many studies do not support such an association. This diversity among the results is due to the heterogeneity of the studies themselves, i.e. the differences in the methodologies used. Interestingly, over the last 20 years numerous researchers have recognised this problem and have identified factors that may induce it. However, in many cases researchers have not taken these factors into account and therefore this diversity has been preserved. This renders any comparison among the studies challenging or even impossible.
The main reason for the heterogeneity among the studies is the inconsistent definition for periodontal disease (gingivitis or periodontitis) and pregnancy complications. Hence, some studies determine the presence or absence of periodontitis by using full-mouth measurements, whereas others use selective measurements such as the Community Periodontal Index of Treatment Needs (CPITN) and Ramfjord index, which may underestimate the severity of periodontal diseases. Moreover, depending upon the study, periodontitis is defined by different clinical parameters, such as probing pocket depth (PPD), clinical attachment level (CAL), bleeding on probing (BOP) and recession. Interestingly, even when the same clinical parameters are used, the definition of periodontitis differs because each research group sets its own threshold values to determine whether a subject has periodontitis or not. Furthermore, some investigators evaluate the association of a pregnancy complication with a clinical parameter such as PPD without providing a specific definition for periodontitis. Finally, some studies included more than one definition for periodontal diseases, depending upon the severity of the disease. To illustrate the importance of choice of case definitions, Manau et al129 calculated the prevalence of periodontal disease to be between 2.2% and 70.8% when 14 different case definitions were applied to the same data from 1296 pregnant women. It is remarkable that although in 2003 the AAP and the United States Centers for Disease Control and Prevention developed standardised clinical case definitions for population-based trials on periodontitis130, and other investigators have also proposed specific definitions of periodontitis for clinical studies131, investigators have not followed unanimously these definitions. Finally, the most common discrepancy in the definition of pregnancy complications is the term PTLBW. Some investigators use this term to describe PTB and LBW, whereas others refer directly to PTB or LBW.
A second reason for the variability among the epidemiological studies is the inadequate adjustment for confounding factors. The age of the pregnant woman, race, socio-economic status, education, smoking, drug use, parity, prenatal care, history of pregnancy complications, genito-urinary (GU) infections, body mass index, oral health and diabetes mellitus are some of risk factors that may affect periodontal health and/or pregnancy outcomes. Therefore these parameters should be adjusted in order to strengthen the validity of the association studies. Unfortunately, the evaluation of the epidemiological studies reveals that a large volume of the available evidence has been generated by studies with poor and inconsistent adjustment of confounding variables. Interestingly, a high racial heterogeneity among the populations studied is evident. Specifically, many studies conducted in the USA led to a positive association between periodontal disease and pregnancy complications, whereas many investigations from Northern Europe did not. This difference may be explained by the fact that a large proportion of the women participating in the USA studies were African American of low socio-economic status, whereas participants of the European studies were mainly white. It is worth noting that APOs in the USA occur more frequently in African-American than in white pregnant women132. Indeed, the incidence of PTB in Europe is 6.2%, in the USA 10.6%, in South America 7.9%, in Australia 6.4%, in Asia 9.1% and in Africa 11.9%. Likewise, the incidence of LBW is 6.4% in Europe, 7.7% in the USA, 9.6% in South America, 6.4% in Australia, 18.3% in Asia and 14.3% in Africa133 (Fig 7-1). Since the prevalence of periodontitis and of pregnancy complications is higher in African Americans than in Caucasians it would be logical to find a stronger association in this group. This should be kept in mind when comparing studies of different ethnic groups and when a generalisation of the conclusions is attempted.
Finally, differences in the size of the study populations exist. For example, in the study published by Agueda et al56 the investigators calculated that for a prospective study with an estimated RR of at least 2, an α error of 5%, and a β error of 10%, more than 1100 subjects should be recruited, considering that the prevalence of periodontal disease in pregnant women is 25% and the prevalence of PTB or LBW is about 7%. However, in many prospective studies a much smaller number of pregnant women (less than 200) have participated, compromising the statistical power of these studies. The same problem is also encountered in case-control studies, where frequently fewer than 20 women have been enrolled in the case and control groups.
As with epidemiological studies, a careful evaluation of the RCTs reveals a significant discrepancy among their designs and execution. One of the most important differences among the RCTs is the diverse definition of periodontal disease used to include pregnant women in the studies and therefore determine whether the exposure is present or not. Therefore, the same population of pregnant women would have been allocated in the exposure group differently based on the various periodontal definitions. Moreover, the extent and severity of periodontal disease, at enrolment, differed among the participants of the studies.
One other critical variation between the RCTs is the effectiveness of the intervention, i.e. to what extent periodontal treatment rendered was able to eliminate the exposure. Four studies did not report the effectiveness of the intervention on periodontal measures106,113,114,134 and half of them did not find a reduction in the rates of APOs after treatment. In the largest RCT112, which showed no effect of periodontal therapy on APOs, although the intervention group had overall better periodontal measurements compared to the control group, periodontal disease progression was reported in 40.7% of the treated women. The lack of effectiveness of the intervention in such high percentage of the intervention group may raise questions concerning the execution of periodontal therapy in this study and may question the value of the results of this RCT. In addition, in two other large RCT studies that also did not show a reduction in the rates of APOs, periodontal therapy significantly reduced periodontal inflammation but not to levels that can be considered as ‘periodontal health’. In the remaining studies, a significant improvement in various clinical periodontal parameters was reported. Since RCTs55,105 that were more effective in reducing periodontal inflammation showed a positive effect in reducing APOs it has been suggested that periodontal treatment efficacy may be important in order to achieve changes in pregnancy outcomes135.
Another variation among the RCTs lies in the type of periodontal treatment rendered. Hence, although in all studies the treatment arm received non-surgical periodontal therapy including oral hygiene instructions (OHI) and scaling and root planing (SRP), not all studies followed a maintenance programme throughout gestation after SRP was completed. In addition, two studies also administered systemic antibiotics (combination of metronidazole and amoxicillin or only metronidazole). Although the beneficial role of these antibiotics in addition to SRP has been demonstrated in non-pregnant patients136,137 it is not clear how the use of metronidazole could have affected the incidence of APOs. Controlled trials show that the use of antibiotics to treat bacterial vaginosis does not reduce the risk of prematurity138, and other studies have shown that oral metronidazole therapy may induce changes in the vaginal flora that are associated with an increased risk of PTB139.
Besides the differences in the intervention groups, in several RCTs women in the control arm also received some kind of intervention ranging from polishing the teeth and OHI to supragingival scaling55,106,108,111,114,134,135. A recent study evaluated an intensive protocol of OHI and dental prophylaxis on pregnant women with gingivitis and found that these procedures decreased gingivitis140. Therefore, it could be suggested that the application of such regimens in the control group would weaken the value of these RCTs since the effect of the intervention on APOs would be counteracted.
One other discrepancy among studies that may explain, in part, the inconsistency in the results of the RCTs, is the variation in the number of participants and the different inclusion criteria regarding APOs and their related risk factors. There are studies that randomised very few pregnant women, i.e. 20 or 30109,118, and others that included more than 700 and up to 1800 subjects107,108,111,112,114. It is clear that the increased sample size gives power to the study. Moreover, given the number of risk factors for APOs, important group imbalances may remain in small trials even with randomisation141. Interestingly, most of the studies showed no effect of periodontal treatment on APOs, which was also reported in a recent umbrella review141a.
Finally, as with the epidemiological studies, inconsistency in the adjustment of confounding factors is present throughout the RCTs. However, in the larger studies that did not find an effect of periodontal treatment on APOs most of the common confounders were controlled141b.
7.2.4 Why periodontal therapy during pregnancy does not seem to affect APOs
Several reasons may explain why non-surgical periodontal therapy during the second trimester of gestation may not have an effect on APOs142. Specifically, one seemingly easy explanation would be that periodontal disease and APOs are not causally linked. Therefore, any periodontal intervention would have little, if any, effect on pregnancy outcomes. However, it should also be borne in mind that the lack of effect of an intervention does not always translate into proof of non-causality. Instead, what it merely shows is that the specific intervention at the time it was rendered was not able to alter the outcome.
Another crucial factor may be the timing of the intervention. Periodontal intervention during the second trimester might have been too late to be able to prevent or even reverse any APO. Perhaps, by the time of treatment, periodontal bacteria may have already reached the foetal-placental unit and may have contributed to the initiation of processes that lead to APOs. In addition, periodontal instrumentation during the second trimester would have induced a transient bacteraemia and elevation of the systemic inflammatory response, which may have further increased the risk for APOs. Indeed, findings from an RCT of the effects of periodontal treatment on vascular endothelial function (a marker of systemic inflammation) demonstrated that periodontal therapy results in immediate, significant impairment of endothelial function that is restored to pre-treatment levels within approximately 2 months after intervention, while the beneficial effects of therapy are manifested at 6 months after intervention143. It is obvious that since periodontal therapy occurs during the second trimester these effects would not be noticeable prior to delivery. Therefore, for all these reasons, it is possible that periodontal therapy during the preconception period may be more meaningful and beneficial for improved pregnancy outcomes.
A third reason that may explain why periodontal treatment during pregnancy does not seem to affect APOs is that in some of the studies periodontal therapy was not effective in improving clinical periodontal parameters to the accepted standard of care. Therefore, an unsuccessful intervention could explain the lack of a reduction in the rates of APOs. Indeed, in studies where periodontal treatment managed to better control gingival inflammation, a positive effect on APOs was noted105,107. Therefore, more strict treatment endpoints may be necessary to have an effect on pregnancy outcomes.
Finally, in many RCTs the women enrolled had very little initial periodontal disease. Hence, in these patients the risk of exposure of the foetal-placental unit to periodontal challenges may have been insignificant even prior to periodontal intervention.
7.2.5 Future research needs
As described earlier, the main reason for more than 100 epidemiological studies having not yet been able to provide clear and strong evidence as to whether periodontal disease is associated with APOs is the large heterogeneity among these studies. Hence, large studies with an appropriate number of participants should be pursued. Moreover, and importantly, the enrolment criteria of the participants should be more universal, allowing periodontally ‘healthy’ and ‘non-healthy’ subjects to be allocated in cases and control groups without selection bias. Similarly, the outcome definitions regarding pregnancy complications should be common. Finally, effort must be made to adjust important confounding factors in order to avoid potential false positive associations.
Regarding the intervention studies, here again, an appropriate number of participants should be selected by using universally accepted periodontal risk exposure thresholds in order to achieve common enrolment criteria. These could refer to clinical, microbiological or serological/immunological parameters. Similarly, common treatment endpoints for each periodontal risk exposure should be used in order to be able to distinguish when periodontal interventions are successful and ‘periodontal health’ is established. The use of the same exposure thresholds and treatment endpoints will render the studies more comparable and will strengthen their conclusions. Unfortunately, to date, these thresholds have not been established and therefore the periodontal community needs to provide some guidelines in this direction. Moreover, alternative treatment modalities such as repeated mechanical debridement or local antimicrobial therapy could also be evaluated. Finally, the establishment of periodontal health in the preconception period and its maintenance during gestation may appear to be more beneficial in reducing the risk of APOs compared to periodontal treatment in the middle of the pregnancy, when the harmful effects of periodontal infection may be already irreversible142.
● There is evidence of a moderate association between periodontal disease and pre-eclampsia. However, this association is weakened by the heterogeneity of the studies.
● There is weak evidence of an association between periodontal disease and PTB or LBW, which is hampered by the extreme heterogeneity of the data.
● Epidemiological studies demonstrate heterogeneity mainly in the populations studied, the means of periodontal assessment and the periodontal disease classification employed, as well as in the way data were handled.
● There is strong evidence that non-surgical periodontal therapy during the second trimester of pregnancy is safe.
● The results from the RCTs are divergent. However, the majority of the higher quality studies do not demonstrate a beneficial effect of non-surgical periodontal therapy during the second trimester on pregnancy outcomes.
● There is limited evidence of a positive effect of periodontal treatment in decreasing PTB and LBW rates in women that are at high risk of APOs.
● No substantial evidence exists regarding the benefit or harm in using systemic antibiotics as adjunct to SRP.
● There is strong evidence that non-surgical periodontal therapy improves the periodontal status of the majority of pregnant women with periodontal disease.
● The heterogeneity among the RCTs is attributed to the characteristics of the populations studied, the sample size, the criteria used for diagnosis of periodontal disease and pregnancy outcomes, the type and effectiveness of the interventions used to control periodontal disease, and the adjustment for confounders.
● It has been suggested that periodontal intervention might be more beneficial in reducing rates of APOs if it occurred during the preconception period. However, no evidence from this kind of intervention is yet available.
7.3.1 Adverse pregnancy outcomes (APOs)
Our understanding of the biology behind periodontal disease and pregnancy complications supported the hypothesis that periodontal infection/inflammation could affect pregnancy outcomes and contribute to pregnancy complications. Therefore, since the first experiments in the early 1990s, numerous mechanistic studies in humans and animals have generated substantial information on the presence of biological pathways that associate periodontal disease with APOs. In order to understand the rationale of these studies and evaluate the importance of their conclusions, a brief description of the aetiopathogenesis of both key players, i.e. periodontal disease and pregnancy complications, is necessary.
220.127.116.11 Aetiopathogenesis of periodontitis
Periodontitis is an infectious disease of the gingiva and of the supporting tissues of the teeth that occurs in a susceptible host. When periodontal microbiota and their virulent by-products interact with the host cells, an immune response is launched to control and eliminate the infection144,145. During this process several signalling molecules (interleukins [IL]-1, -8 and -6, tumour necrosis factor alpha [TNF-α], inflammatory mediators [prostaglandin E2 (PGE2)], and enzymes related to tissue destruction [i.e. matrix metalloproteinases (MMPs)]) are produced locally in the periodontium146. The subsequent tissue breakdown and bleeding allow these inflammatory mediators and the bacteria and their by-products to spill in the blood circulation. This induces a transient bacteraemia and the activation of a systemic inflammatory response147. Thus, acute phase response proteins such as C-reactive protein (CRP) and fibrinogen are produced by the liver and are released back to the systemic circulation through which they can reach areas where inflammation is present and enhance the inflammatory processes.
18.104.22.168 Physiology of normal pregnancy
The placenta is a vessel-rich organ that provides favourable conditions for foetal growth by permitting gas exchange, nutrient uptake and waste disposal through the mother’s blood circulation. This transportation is enabled via the umbilical cord, which connects the foetus to the placenta. Having the necessary resources, the foetus grows in the amniotic fluid, which is contained by the amniotic sac. The walls of this sac consist of two layers, the amnion and the chorion. They are, like the placenta, attached to the uterus through the decidua and the myometrium. As the foetus grows the increasing needs for nutrients and the decreasing space become critical parameters for the survival of both mother and foetus. Hence, as pregnancy progresses, amniotic fluid levels of PGE2 and inflammatory cytokines, such as TNF-α, IL-1β and MMPs rise steadily until a critical threshold level is reached to induce rupture of the amniotic sac membranes, uterine contraction, cervical dilation and delivery148,149. Thus, normal parturition is controlled by inflammatory signalling and this process represents a triggering mechanism that can be modified by external stimuli including infection and inflammatory stressors150. In addition, during gestation the maternal immune system has to defend both the mother and the foetus from external pathogens, but simultaneously it needs to be suppressed in order to tolerate the foetal components inherited from the father151. Thus, in successful gestation, a trend from T-helper (Th)1 towards the Th2 cytokine profile seems to occur both in the peripheral blood and at the foetal-maternal interface152. In brief, Th1 cytokines are macrophage-activating effector molecules while Th2 cytokines are more B-cell-activating effector molecules.
22.214.171.124 Aetiopathogenesis of APOs
As described above, the presence of inflammatory mediators is necessary for normal parturition. However, excessive accumulation of these mediators within the uterus may become dangerous for the mother and the foetus and are associated with various pregnancy complications. These inflammatory mediators may be produced locally or may arrive from the systemic circulation of the mother. Interestingly, although these mediators reach the placenta not all them will pass the placental barrier and reach the foetus. In general, larger proteins (> 500 kDa) are not transferred through the placenta. Regarding the inflammatory cytokines, Il-6 may cross the barrier, while IL-1α, TNF-a and CRP do not153,154. Prematurely elevated levels of IL-1β, IL-6, TNF-α, PGE2, fibronectin and α-fetoprotein and often of MMPs in the amniotic fluid have been associated with PTB by providing the necessary signalling to initiate processes such as uterine contraction and preterm rupture of the membranes, leading to premature delivery152,155–158. Moreover, increased maternal serum levels of pro-inflammatory cytokines, such as IL-1, IL-6, IL-8 and TNF-α have also been reported to be associated with prematurity or low birth weight (and PTLBW)156,159–163.
Similarly, increased concentrations of markers of systemic inflammation, such as CRP, have also been related with PTB162. Interestingly, it has been suggested that the enhanced levels of CRP in maternal serum indicates the presence of another inflammatory condition, i.e. infection156. Besides PTB, elevated levels of CRP have also been linked with an increased risk for other pregnancy complications, such as intrauterine growth restriction (IUGR)163 and pre-eclampsia164. Although an imbalance between angiogenic and anti-angiogenic factors has emerged as a central pathogenic mechanism of pre-eclampsia165, high levels of pro-inflammatory cytokines have also been observed not only in the peripheral blood but also in the placenta and umbilical cord166. Finally, elevated serum CRP, IL-6 and TNF-α in women with gestational diabetes mellitus (GDM) suggests a role for inflammation in the aetiology of this pregnancy complication. It is known that IL-6 and TNF-α interfere with insulin signalling, and are also insulin antagonists. Therefore, sustained elevated levels of TNF-α during late pregnancy are inversely correlated with insulin sensitivity that can result in GDM167.
The enhanced inflammatory response in the foetal-placental unit that occurs in cases of APOs could be partially justified by the presence of intrauterine infections168–170. Microorganisms can invade the intrauterine environment and gain access to various sites of the foetal-placental compartment, such as the choriodecidual space (deciduitis), the chorioamniotic membrane (chorioamnionitis), the placenta (villitis), the amniotic fluid, the umbilical cord (funisitis) or the foetus (sepsis)59,81,171. Therefore, the womb does not provide a protective barrier against microbial invasion and should not be considered a sterile environment. Microorganisms can reach the intrauterine space either by ascending from the vagina and the cervix, in cases of GU tract infections, or by haematogenous dissemination through the placenta. However, accidental introduction of microorganisms in the amniotic cavity may also occur during invasive procedures such as amniocentesis3.
During these infections immune cells from the mother and the foetus are recruited to the area and the inflammatory processes that are initiated may induce tissue damage. For example, in chronic chorioamnionitis, maternal CD8+ T cells infiltrate the amniotic sac membranes, induce trophoblast apoptosis, and damage the chorionamniotic membranes172. This represents the most common lesion in late spontaneous PTB and the frequency is remarkably high, not only in preterm rupture of membranes (39%) and preterm delivery (34%), but also in terms of foetal death (60%)172. Interestingly, the percentage of APOs due to intrauterine infections may be even higher due to difficulties in culturing techniques and the presence of subclinical infections3,168,173.
The role of urinary and vaginal infections in the development of various APOs has also been demonstrated indirectly by studies that used systemic antibiotics such as metronidazole and clindamycin in order to control these infections early during pregnancy. Although not always consistent, these studies show a decrease in the incidence of PTB, LBW and pre-eclampsia after antibiotic treatment174–176. Finally, several non-GU tract infections, such as pyelonephritis, asymptomatic bacteriuria, pneumonia, and appendicitis have been associated with, and probably predispose to PTB177,178.
● In order for periodontal disease (gingivitis or periodontitis) to occur, the host immune response has to be dysregulated or/and there has to be a shift towards a more pathogenic microbial flora.
● Normal parturition is controlled by inflammatory signalling and this process represents a triggering mechanism that can be modified by external stimuli including infection and inflammatory stressors.
● Excessive accumulation of inflammatory mediators within the uterus is associated with various pregnancy complications.
● Intrauterine infections have been associated with APOs.