Periodontal disease and adverse pregnancy outcomes: is there an association?

Article Outline

• Abstract
• Materials and Methods
  • Statistical analysis
  • Sample size
• Results
• Comment
• References
• Copyright

Objective

We assessed the risk of adverse pregnancy outcomes (preterm birth [PTB], preeclampsia [PRE], fetal growth restriction [FGR], or perinatal death) in women with periodontal disease (PD) compared to those without.

Study Design

A multicenter prospective cohort study enrolled women from 3 sites between 6 and 20 weeks' gestation. The presence of PD was defined as periodontal attachment loss ≥ to 3 mm on 3 or more teeth. The primary binary composite outcome included PRE, PTB, FGR, or perinatal death. Multivariable logistic regression (MVLR) was used to control for confounders.

Results

Three hundred eleven patients with and 475 without PD were included. There was no association between PD and the composite outcome, PRE, or PTB in unadjusted analyses. There was no association between PD and the composite outcome (adjusted odds ratio [AOR], 0.81; 95% confidence interval [CI], 0.58-1.15; P = .24), preeclampsia (AOR, 0.71; 95% CI, 0.37-1.36; P = .30), or preterm birth (AOR, 0.77; 95% CI, 0.49-1.21; P = .25) after adjusting for relevant confounders.

Conclusion

Despite the body of literature suggesting an association between PD and adverse pregnancy outcomes in urban populations, this large prospective study failed to demonstrate an association.

Key words: adverse outcomes, periodontal disease, pregnancy, preterm birth

There are 4 million deliveries every year in the United States.¹ Approximately 12% are affected by preterm delivery and 5-7% are affected by preeclampsia.¹, ² Preeclampsia is a major contributor to maternal and perinatal morbidity and mortality. Stillbirth and growth restriction are additional adverse pregnancy outcomes that affect many women each year and are difficult to reliably predict. Despite advances in technology, promotion of prenatal care and continued scientific efforts focused on reducing adverse pregnancy outcomes, little reduction has occurred. This is largely due to an incomplete understanding of the etiology of these adverse outcomes. Infection and/or inflammation as a causative factor for these adverse outcomes continue to be at the forefront of etiologic theories. Previous studies have demonstrated a link between infection and/or inflammation and preterm birth, preeclampsia, and other adverse outcomes thought to be in part secondary to poor placentation.³, ⁴, ⁵, ⁶ However, no successful targeted interventions have been developed likely due to the lack of clear inciting agents.

Given the theorized link between infection and/or inflammation and many adverse pregnancy outcomes, it is biologically plausible that periodontal disease may be linked to adverse pregnancy outcomes. Periodontal disease is one of the most common chronic infectious disorders in humans with prevalence between 10 and 60% depending on the definition and the population being studied.⁷, ⁸ Within the past decade, several studies have demonstrated an association between periodontal disease and the development of systemic diseases such as atherosclerosis and diabetes.⁹, ¹⁰ The underlying theory is that periodontal disease leads to a chronic systemic inflammatory response that then influences the onset and course of cardiovascular disease and diabetes mellitus. Adverse pregnancy outcomes including preeclampsia, low birthweight, and preterm delivery have also been among the “systemic diseases” studied in association with periodontal disease.¹⁰, ¹¹ In addition to biologic plausibility, to further support this possible link, periodontal disease occurs commonly in the general population, but does have an increased prevalence in underserved populations. These are the same populations disproportionately affected by adverse pregnancy outcomes.

Studies performed to date examining the relationship between periodontal disease and pregnancy outcomes have mostly focused on the relationship between periodontal disease and a single pregnancy outcome, including preterm birth, preterm low birthweight, and low birthweight. Some have demonstrated a positive association,¹², ¹³, ¹⁴, ¹⁵, ¹⁶, ¹⁷, ¹⁸, ¹⁹ while others have failed to demonstrate an association.¹², ²⁰, ²¹, ²², ²³, ²⁴, ²⁵ Interestingly, many of the studies that have failed to demonstrate an association between periodontal disease and adverse outcomes have been performed in other countries (United Kingdom, Canada, Denmark, and Sri Lanka). Studies that have shown a positive association between periodontal disease and various adverse outcomes have been performed in the United States, in largely urban settings with high proportions of African American patients. This suggests that periodontal disease may have varying associations with adverse pregnancy outcomes in different populations, the etiology of periodontal disease varies in different populations, or that periodontal disease is a marker for other factors that are associated with adverse pregnancy outcomes.

To this end, we sought to compare the risk of a composite of adverse pregnancy outcomes in women with and without prospectively identified (early in pregnancy) periodontal disease in a multicenter urban population. Our secondary aims were to compare the risk of specific adverse outcomes, preeclampsia and preterm birth, in women with and without periodontal disease.

Back to Article Outline

Materials and Methods 

The Periodontal Infection and Prematurity Study (PIPS) includes a large, randomized controlled trial to study the association between periodontal disease, treatment, and preterm birth. The primary aim of the overall randomized trial is to compare efficacy of scaling and planning treatment of periodontal disease to polishing (placebo) in preterm birth prevention (< 35 week delivery). Additionally, an observational cohort of women without periodontal disease was followed as a comparison group.

We performed a multicenter prospective cohort study. Specifically, we set out to compare women who screened negative for periodontal disease (unexposed group), and were thus ineligible for the trial, to those who screened positive for periodontal disease (exposed group), but were not randomized, because they failed to show up for the follow-up visit when randomization occurred.

Patients for the PIPS study were recruited from 3 centers in Philadelphia: Helen O. Dickens Center at the Hospital of the University of Pennsylvania (HUP), Women and Children's Health Service at Pennsylvania Hospital (PAH), and Albert Einstein Medical Center (AE). Recruitment occurred from October 2004 to October 2007. All of these sites provide prenatal care for underserved populations in Philadelphia, and all enrollment occurred in large prenatal clinics at each site. Trained research nurses from the 3 participating sites enrolled patients for the PIPS trial. Ninety percent of women who deliver at these hospitals enroll in prenatal care at < 20 weeks' gestation at these 3 sites based on a prior study performed by one of the investigators (data not shown).

Inclusion criteria for the PIPS study (randomized and observational components) included women who were 6-20 weeks' gestation (adjusted by ultrasound if necessary). A nurse trained by dental personnel evaluated women for the presence or absence of periodontal disease using predefined criteria. Each study nurse underwent retraining and observation once to twice annually for the duration of the study. A thorough, complete exam was performed on all patients. Specifically, the presence of periodontal disease was defined as periodontal attachment loss greater than or equal to 3 mm on 3 or more teeth. Eligibility for enrollment was determined by presence or absence of periodontal disease (dichotomous yes/no); those with documented periodontal disease were eligible for the randomized trial. Bleeding on probing and severity of periodontal disease at initial screening were not recorded in this study. Those without periodontal disease were eligible to be followed in the observational cohort. Women who received periodontal treatment during the pregnancy, used antibiotic or antimicrobial mouthwash within 2 weeks of enrollment, had a multiple gestation pregnancy, or had known mitral valve prolapse were excluded.

Women identified with periodontal disease were asked to complete a baseline interview and provide consent for enrollment into the trial. Randomization took place at their subsequent visit. Women with periodontal disease who were enrolled initially but who failed to present for randomization formed the exposed cohort in this investigation. Women without documented periodontal disease at the initial visit were asked to complete a baseline interview and provide consent for enrollment into a follow-up study. These women formed the unexposed cohort in this investigation. All data were collected by a structured interview at enrollment and chart review by trained research nurses at the end of the follow-up period. Data collected from these 2 sources included outcomes as described below as well as potential identifiable confounders. Potential confounders that were analyzed include: prior obstetric history, maternal race, age, and socioeconomic status (measured by highest level of educational attainment), site of care, obesity, diabetes and chronic hypertension (prior to pregnancy), tobacco use, drug use, sexually transmitted infection, and bacterial vaginosis. Body mass index (BMI) was determined using height and current weight given by patient at time of interview. Periodontal disease status was not reassessed at the time of delivery in any study group.

The primary study outcome was a binary composite outcome and included the occurrence of any of the following 4 outcomes: preterm birth (spontaneous or indicated delivery at < 37 weeks' gestation), preeclampsia (hypertension after 20 weeks' gestation based on traditional criteria per American College of Obstetricians and Gynecologists [≥ 140/90 mmHg] with proteinuria),²⁶ fetal growth restriction (< 10% for gestational age based on Alexander growth curve [FGR]),²⁷ or perinatal death (20 weeks' gestation to first 7 days of life). Biologically, it is postulated that these outcomes are all linked with the common possible etiology of inflammation, which provides a rationale for grouping these outcomes as a composite outcome. Trained research nurses obtained all outcome information from hospital chart abstraction after delivery. There was minimal loss to follow-up since only a small proportion of patients (≲ 5%) did not deliver at their enrollment site of care.

Statistical analysis 

Descriptive statistics (means, median, frequencies) were calculated to describe the data, evaluate the completeness of the data collected, and formulate rules to categorize variables for further analyses. Continuous variables were compared using Student t test. Categorical or dichotomous variables were compared using χ² analysis or Fisher exact test. To compare women with and without periodontal disease for the composite outcome, preeclampsia and preterm birth (< 37 weeks), we performed an initial bivariate analysis with Fisher exact test and χ² test, depending on the available sample size for each cell. For this cohort study, all unadjusted analyses are reported as relative risks. Stratified analyses by site, race, and history of preterm delivery were performed to assess potential effect modification and confounding. Multivariable logistic regression was then used to control for confounding. All adjusted analyses are reported as odds ratios, which approximate the relative risks when the rare disease assumption is met as it is in the case of the adverse outcomes reported in this study.

In all models, the composite adverse pregnancy outcome or individual secondary adverse outcome was the dependent variable and periodontal disease was an independent covariate. Pre-specified potential confounders were added to the model based on biologic plausibility and statistical significance in the unadjusted analysis. Covariates were retained in the model if they changed the effect size around the primary covariate by > 10%. Site was included in all final adjusted models as a categorical variable. Body mass index (BMI) was included as a dichotomous variable “obesity” ≥ 30 or < 30. Maternal age was examined as a categorical variable (< 20, 21-34, ≥ 35 years).

Sample size 

The outcome of preeclampsia was used to calculate the sample size since this is the least prevalent outcome among our aims (composite outcome, preeclampsia, and preterm birth). Assuming a ratio of unexposed to exposed 1.5:1, a type I, alpha, error of 0.05 (2-sided), 80% power, and a cumulative incidence of preeclampsia of 5% in the unexposed group, 305 patients with periodontal disease and 461 patients without periodontal disease were required to detect a minimal relative risk increase of 2.15.

Back to Article Outline

Results 

In the overall study, 5085 women were screened at the 3 study sites between October 2004 and October 2007. Six hundred ninety-five declined participation, 927 women were excluded based on the predetermined exclusion criteria, and 352 women never had a periodontal evaluation. Trained nurses screened 3111 women for periodontal disease. One thousand five hundred and sixty-six screened positive (50.3%) and 1545 women (49.7%) screened negative. Of these, 311 (20% of screen positive patients) women with periodontal disease and 475 (30.7% of screen negative patients) women without periodontal disease were included in this cohort study. At the time of this cohort study, deliveries had occurred and pregnancy outcomes were available for 83% of those with periodontal disease and 80% of those without (Figure 1).

• 
  • View Large Image
  • Download to PowerPoint

• FIGURE 1. 

  Patient enrollment flow diagram

• This figure demonstrates the flow of patients from screening for periodontal disease through inclusion in our prospective observational study.

• Srinivas. Periodontal disease and adverse pregnancy outcomes. Am J Obstet Gynecol 2009.

Table 1 demonstrates the overall demographic characteristics of the cohort comparing the exposed to the unexposed groups. The percent exposed and unexposed were slightly different among the sites. Women with periodontal disease screened at a slightly greater gestational age (both in the first trimester), were older, had a greater maternal weight, were more likely to be African American, and use tobacco. These are all known risk factors for periodontal disease. The mean maternal age was 23.9 years in the exposed group compared to 22.3 years in the unexposed group (P < .001). The mean maternal weight was 173.3 pounds in the exposed group compared to 164.7 in the unexposed group (P = .02). There was no difference in level of education, or prevalence of women with chronic hypertension, diabetes, bacterial vaginosis, or history of a prior preterm delivery between the groups. Neither the mean gestational age at delivery nor the mean birthweight was different between the 2 groups. The mean gestational age at delivery was 38.2 weeks and 38.3 weeks in the exposed and unexposed groups, respectively (P = .48). The mean birthweight in the exposed group was 3096 g compared to 3129 g in the unexposed group (P = .48).

TABLE 1. Demographic characteristics between women with and without periodontal disease

Characteristic	Periodontal disease % (n = 311)	No periodontal disease % (n = 475)	P value
Site			<.001
Site 1 (HUP)	44(137)	28.6(136)
Site 2 (PAH)	42.8(133)	57.2(272)
Site 3 (AE)	13.2(41)	14.1(67)
Gestational age at screena	13±3.4(13)	12.3±3.6(12)	.01
Education (≤ HS, GED)	69.8(211)	60.3(291)	.13
Race (African American)	86.5(269)	77(366)	.002
Obese (BMI ≥ 30)	43.1(134)	30.1(143)	<.001
Chronic hypertension	6.5(20)	6.78(32)	.9
Diabetes	1.6(5)	2.8(13)	.3
Current tobacco use	16(49)	10.2(48)	.02
History of bacterial vaginosis	30.6(93)	31.6(149)	.77
History of preterm delivery	11.6(36)	13.3(63)	.49

AE, Albert Einstein Medical Center; BMI, body mass index; GED, general educational development; HS, high school; HUP, Helen O. Dickens Center at the Hospital of the University of Pennsylvania; PAH, Women and Children's Health Service at Pennsylvania Hospital.

Srinivas. Periodontal disease and adverse pregnancy outcomes. Am J Obstet Gynecol 2009.

Table 2 provides the prevalence and the unadjusted relative risks for the composite outcome and the individual components in the exposed group compared to the unexposed. There was no association between periodontal disease and the composite outcome or any of the individual components. When spontaneous and indicated preterm birth at < 37 weeks were examined separately, there was no association between either of these outcomes and periodontal disease (spontaneous: 9.6% in exposed and 11.2% in unexposed, P = .47; indicated: 2.8% in exposed and 4.9% in unexposed, P = .17). Further, when preterm birth < 35 weeks was evaluated, there was no difference in this outcome between those with periodontal disease (7.1%) and those without (7.2%, P = .96). Additionally, when the association between mild and severe preeclampsia and periodontal disease were evaluated separately, there were no associations seen (mild: 1.3% in exposed and 2.1% in unexposed, P = .4; severe: 3.9% in exposed and 4.6% in unexposed, P = .6). We also performed stratified analyses by maternal age, race, site of care, gestational age at screen, chronic hypertension, diabetes, obesity, and history of preterm delivery to assess for potential effect modification and confounding. There was no evidence of effect modification by any of these covariates (data available upon request).

TABLE 2. Outcomes and unadjusted relative risks in women with and without periodontal disease

Outcome	Percent with periodontal disease % (n)	Percent without periodontal disease % (n)	RR	95% CI	P value
Composite adverse outcome	23.8(74)	28(134)	0.87	0.71-1.07	.17
Preeclampsia	5.2(16)	6.7(32)	0.84	0.55-1.26	.37
Preterm birth (< 37 wk)	11.9(37)	15.2(72)	0.84	0.64-1.11	.20
FGR	12.5(39)	15.4(73)	0.86	0.66-1.13	.27
Perinatal death	0.3(1)	0.4(2)	0.84	0.17-4.19	.83

CI, confidence interval; FGR, fetal growth restriction.

Srinivas. Periodontal disease and adverse pregnancy outcomes. Am J Obstet Gynecol 2009.

Although no significant association was seen between periodontal disease and the adverse composite outcome, preeclampsia or preterm birth, adjusted analyses were performed to control for the presence of known confounders (Table 3). In our final multivariable logistic regression model for the composite outcome, there was no association between periodontal disease and the composite outcome (adjusted odds ratio [AOR], 0.81; 95% confidence interval [CI], 0.58-1.15; P = .24) after adjusting for site, maternal age, race, tobacco, and obesity. Further, there was no association between periodontal disease and preeclampsia after adjusting for site, maternal race, age, tobacco, obesity, and chronic hypertension. In these adjusted analyses, chronic hypertension was associated with preeclampsia (AOR, 3.54; 95% CI, 1.48-8.48; P = .005). There was also no association between periodontal disease and preterm birth after adjusting for site, maternal age, race, tobacco, obesity, and history of a prior preterm delivery. History of a prior preterm delivery was associated with preterm delivery (AOR, 3.24; 95% CI, 1.88-5.58; P < .001), as well as maternal age less than 20 years old (AOR, 1.62; 95% CI, 1.0-2.61; P = .05).

TABLE 3. Adjusted analysis for adverse composite outcome, preeclampsia, and preterm birth

Variable	AOR	95 % CI	P value
Composite outcomea
Periodontal disease	0.81	0.58-1.15	.25
Preeclampsiab
Periodontal disease	0.71	0.37-1.36	.30
Chronic hypertension	3.54	1.48-8.48	.005
Preterm birthc
Periodontal disease	0.77	0.49-1.21	.25
History of preterm delivery	3.24	1.88-5.58	<.001

AOR, adjusted odds ratio; CI, confidence interval.

Srinivas. Periodontal disease and adverse pregnancy outcomes. Am J Obstet Gynecol 2009.

Additionally, to exclude the possibility that our findings represent a selection bias, we performed a planned comparison between exposed women who were not randomized (the exposed group in this cohort study) to exposed women who were randomized in the PIPS trial. Table 4 demonstrates a comparison between these 2 groups. There is a minimal difference in gestational age at screen between the 2 groups, although the mean screening age is in the first trimester for both groups. There is also a trend towards more patients with diabetes in the randomized group. There are no other significant differences between these 2 exposed groups.

TABLE 4. Women with periodontal disease: comparison of patients not randomized in trial to patients in randomized trial

Characteristic	Periodontal disease—not randomized % (n = 311)	Periodontal disease—randomized % (n = 658)	P value
Site			<.0010
Site 1 (HUP)	44(137)	56.4(371)
Site 2 (PAH)	42.8(133)	22.8(150)
Site 3 (AE)	13.2(41)	20.8(137)
Gestational age at screena	13±3.4(13)	12.5±3.5(12)	.04
Mean maternal agea	23.9±4.7(23)	24.3±5.5(23)	.19
Education (≤ HS, GED)	69.78(217)	69.8(459)	.85
Race (African American)	86.5(269)	577(87.7)	.6
Mean maternal weighta	173±47(166)	173±51(162)	.93
Obesity	43.1(134)	37.7(248)	.11
Chronic hypertension	6.5(20)	8.7(57)	.25
Diabetes	1.6(5)	3.7(24)	.09
Current tobacco use	16(49)	13(86)	.23
History of bacterial vaginosis	30.6(93)	28.0(184)	.47
History preterm delivery	11.6(36)	13.5(89)	.40
Mean gestational age at delivery (wk)a	38.3±2.4(39)	37.9±3.6(39)	.06
Mean birthweighta	3129±602(3197)	3055±730(3175)	.12

AE, Albert Einstein Medical Center; BMI, body mass index; GED, general educational development; HS, high school; HUP, Helen O. Dickens Center at the Hospital of the University of Pennsylvania; PAH, Women and Children's Health Service at Pennsylvania Hospital.

Srinivas. Periodontal disease and adverse pregnancy outcomes. Am J Obstet Gynecol 2009.

Back to Article Outline

Comment 

In our multicenter study, we did not observe an association between the presence of periodontal disease and a composite adverse pregnancy outcome that included preeclampsia, preterm birth, intrauterine growth restriction, and perinatal death. In addition, there were no significant observed associations between periodontal disease and the individual outcomes of preeclampsia and preterm birth.

These findings are in contrast to several observational studies that have demonstrated an association between periodontal disease and a variety of adverse outcomes including preeclampsia, preterm birth, and low birthweight.¹³, ¹⁴, ¹⁵, ¹⁶, ¹⁷, ¹⁸, ¹⁹ However, our findings are consistent with several international trials as well as with findings of a recent large multicenter randomized trial in the United States that evaluated the efficacy of treatment of periodontal disease in reducing preterm birth.²⁰, ²¹, ²², ²³, ²⁴, ²⁵, ²⁸

There is a significant body of literature suggesting the role of infection/inflammation in adverse pregnancy outcomes. Thus, the relationship between periodontal disease and adverse pregnancy outcomes is biologically plausible. However, there are several potential explanations for the inconsistencies in the literature regarding the association between periodontal disease and adverse pregnancy outcomes. First, there are several varying criteria used to define periodontal disease in the literature, and the timing of this assessment in relation to pregnancy outcome is inconsistent. A recent study in the Journal of Clinical Periodontology applied a variety of different published periodontal disease definitions to a large cohort study about periodontal disease performed in Spain. They found that the association between periodontal disease and adverse outcomes varied based on the definition utilized.²⁹ Further, some retrospective studies that have found an association between periodontal disease and adverse outcomes assessed the exposure of periodontal disease post delivery, thus raising the question of whether the exposure was truly present in early pregnancy prior to the outcome.¹⁹ A third explanation may be due to the effect of confounders. Periodontal disease is more common in economically disadvantaged populations, the same women who are at increased risk for adverse pregnancy outcomes. The studies that have consistently demonstrated an association between periodontal disease and a variety of pregnancy outcomes have been largely urban populations, whereas those with consistent negative associations have been largely international. The finding of an association in mixed urban populations in the United States may reflect the fact that periodontal disease or poor dentition may be a marker of other factors that are associated with adverse pregnancy outcomes (low socioeconomic status, poor access to health care, environmental exposures, nutrition) and not periodontal disease itself. This multicenter cohort study in an urban United States population further strengthens the existing body of literature that demonstrates no association between periodontal disease and adverse pregnancy outcomes. Specifically, almost 90% of our group with periodontal disease was African American, and almost 70% had a high school education or less. This suggests, along with our negative findings, and negative findings in other international population studies, that periodontal disease itself is not associated with adverse outcomes. The positive findings in other urban populations may be due to a more socio-economically mixed population and that, in those studies, periodontal disease was actually a marker for low socioeconomic status.

There are several strengths to this study. Importantly, this was a cohort study within a randomized trial. Inclusion and exclusion criteria were strict and prespecified. The exposure (periodontal disease) was prospectively assessed by trained personnel using strict criteria determined by collaborators in dentistry, thus minimizing the chance for misclassification of the exposure. The low chance of misclassification is further demonstrated by the agreement between screening diagnosis of periodontal disease and repeat examination by a dental hygienist at time of enrollment into the treatment trial (> 98%). Further, in the patients randomized, there was 99% correlation between the measures of periodontal screening and the measurements performed at the time of detailed dental hygienist exam at randomization. The prospective identification of the exposure is an important advantage of this study compared to previous studies. Further, it was multicenter with a heterogeneous population, making the findings more generalizable. Additionally, a large amount of clinical data was collected on all patients, thus allowing important confounders to be included in our final regression models. Finally, the study had a large number of exposed and unexposed patients, thus providing statistical power to assess a clinically meaningful difference in the composite outcome, preeclampsia and preterm birth.

However, this design does have some limitations that should be acknowledged. First, the outcome preeclampsia was identified by chart abstraction based on individual clinicians. This may result in misclassification of patients as having the outcome when they truly did not, or vice versa. If this misclassification did occur, it is likely to be nondifferential since the exposure status was unknown to clinicians, and would bias our findings toward the null. However, this is not likely to have occurred since all of the participating centers use the standard American College of Obstetricians and Gynecologists definition for preeclampsia in routine practice.²⁶ Further, the outcome perinatal death may miss neonates that were discharged alive and died at home. However, we do not anticipate this to be a significant problem since this is an extremely unusual occurrence within the first week of life. There is also a potential for selection bias relative to nonparticipation. Although all women screened do obtain prenatal care at one of the 3 clinical sites, the women in our exposed cohort are those who did not return for a follow-up visit to be randomized, raising the concern that these women may be different at baseline from the women who are randomized. When these 2 groups were compared (the exposed in this cohort study vs the exposed women who were randomized), they were similar with respect to almost all demographic characteristics. This strengthens our study findings and demonstrates that the exposed cohort in this study is similar to the exposed women who were randomized in the trial. Additionally, since the women included in this cohort study were not randomized, the exposure of periodontal disease is only characterized as a dichotomous variable, and we lack additional measures of disease severity in this group. There is also the possibility of unmeasured confounding. We have included a comprehensive list of possible confounders; however, there may be factors that affect both the exposure and the outcome that are either unknown or un-measurable. Further, the reported preterm birth rate for non-Hispanic African Americans in Pennsylvania on average is 16.8%. This rate includes patients who did not receive prenatal care. In our multicenter urban population, the preterm birth rate is 11.9% in those with periodontal disease and 15.2% in those without, despite the large proportion of African Americans included. In the overall trial, the preterm delivery rate for non-Hispanic African Americans was 14.7% in those who were randomized. All women in this study initiated prenatal care at < 20 weeks' gestation, and this may largely account for the lower preterm birth rate observed. This type of nondifferential selection bias, in which healthier patients tend to volunteer for study enrollment, is unlikely to produce an exaggerated estimate of effect size (relative risk), and has more predictable affects on results, generally biasing estimates toward the null hypothesis. Importantly, the possibility of nondifferential selection bias does not invalidate our study results, particularly since we observed a lower, rather than the anticipated higher, preterm birth rate in the patients with periodontal disease. Finally, we cannot exclude a small association between the exposure of periodontal disease and our outcomes. However, given our sample size and the 95% confidence interval around the point estimates for each association between exposure and outcome, this is unlikely. If there is a true positive association, it is likely small and clinically not significant.

Despite the body of literature suggesting an association between periodontal disease and adverse pregnancy outcomes in urban populations, this study failed to replicate that association. Although the biologic plausibility and the potential for a treatment that could reduce preterm birth and other adverse outcomes were promising, this study demonstrates that no association between periodontal disease and adverse pregnancy outcomes exists. Continued research to investigate causes of adverse pregnancy outcomes and possible prevention strategies should be pursued.

Back to Article Outline

References 

1. Centers for Disease Control and Prevention. CDC statistics 1991-1999.
   • View In Article
2. Sibai B, Dekker G, Kupferminc M. Preeclampsia. Lancet. 2005;365:785–799
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (141 KB)
   • CrossRef
3. Gibbs RS, Romero R, Hillier SL, et al. A review of premature birth and subclinical infection. Am J Obstet Gynecol. 1992;166:1515–1528
   • View In Article
   • Abstract
   • Full-Text PDF (2189 KB)
4. Gibbs RS. The relationship between infections and adverse pregnancy outcomes: an overview. Ann Periodontol. 2001;6:153–163
   • View In Article
   • MEDLINE
   • CrossRef
5. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet. 2008;371:75–84
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (378 KB)
   • CrossRef
6. Sibai B, Dekker G, Kupferminc M. Preeclampsia. Lancet. 2005;365:785–799
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (141 KB)
   • CrossRef
7. Albandar JM, Rams TE. Global epidemiology of periodontal diseases: an overview. Peridontol 2000. 2002;29:7–10
   • View In Article
8. Papapanou PN. Periodontal diseases: epidemiology. Ann Periodontol. 1996;1:1–36
   • View In Article
   • MEDLINE
   • CrossRef
9. Garcia RI, Henshaw MM, Krall EA. Relationship between periodontal disease and systemic health. Periodontol 2000. 2001;25:21–36
   • View In Article
   • MEDLINE
   • CrossRef
10. Paquette DW. The periodontal infection-systemic disease link: a review of the truth or myth. J Int Acad Periodontol. 2002;4:101–109
    • View In Article
    • MEDLINE
11. Xiong X, Buekens P, Fraser WD, Beck J, Offenbacher S. Periodontal disease and adverse pregnancy outcomes: a systematic review. BJOG. 2006;113:135–143
    • View In Article
    • MEDLINE
    • CrossRef
12. Xiong X, Buekens P, Vastardis S, Yu SM. Periodontal disease and pregnancy outcomes: state of the science. Obstet Gynecol Surv. 2007;62:605–615
    • View In Article
    • CrossRef
13. Jeffcoat MK, Geurs NC, Reddy MS, Cliver SP, Goldenberg RL, Hauth JC. Periodontal infection and preterm birth. J Am Dent Assoc. 2001;132:875–888
    • View In Article
    • MEDLINE
14. Offenbacher S, Katz V, Fertik G, et al. Periodontal disease as a possible risk factor for preterm low birth weight. J Periodontol. 1996;67:1103–1113
    • View In Article
    • MEDLINE
    • CrossRef
15. Mokeem SA, Molla GN, Al-Jewair TS. The prevalence and relationship between periodontal disease and pre-term low birth weight infants at King Khalid University Hospital in Riyadh, Saudi Arabia. J Contemp Dent Practice. 2004;5:40–56
    • View In Article
16. Goepfert AR, Jeffcoat MK, Andrews WW, et al. Periodontal disease and upper genital tract inflammation in early spontaneous preterm birth. Obstet Gynecol. 2004;104:777–783
    • View In Article
    • MEDLINE
    • CrossRef
17. Dasanayake AP. Poor periodontal health of the pregnant woman as a risk factor for low birth weight. Ann Peridontol. 1998;3:206–212
    • View In Article
18. Radnai M, Gorzo I, Nagy E, Urban E, Novak T, Pal A. A possible association between preterm birth and early periodontitis (Pilot study). J Clin Periodontol. 2004;31:736–741
    • View In Article
    • MEDLINE
    • CrossRef
19. Boggess KA, Lieff S, Murtha AP, Moss K, Beck J, Offenbacher S. Maternal periodontal disease is associated with an increased risk for preeclampsia. Obstet Gynecol. 2003;101:227–231
    • View In Article
    • MEDLINE
    • CrossRef
20. Moore S, Ide M, Coward PY, et al. A prospective study to investigate the relationship between periodontal disease and adverse pregnancy outcome. Br J Dent. 2004;197:251–258
    • View In Article
21. Davenport ES, Williams CE, Sterne JA, Murad S, Sivapathasundram V, Curtis MA. Maternal periodontal disease and preterm low birth weight: case-control study. J Dent Res. 2002;81:313–318
    • View In Article
    • MEDLINE
    • CrossRef
22. Skuldbol T, Johansen KH, Dahlen G, et al. Is preterm labor associated with periodontitis in a Danish maternity ward?. J Clin Periodontol. 2006;33:177–183
    • View In Article
    • MEDLINE
    • CrossRef
23. Rajapakse PS, Nagarathne M, Chandrasekra KB, et al. Periodontal disease and prematurity among non-smoking Sri Lankan women. J Dent Res. 2005;84:274–277
    • View In Article
    • MEDLINE
    • CrossRef
24. Moore S, Randhawa M, Ide M. A case-control study to investigate an association between adverse pregnancy outcome and periodontal disease. J Clin Periodontol. 2005;32:1–5
    • View In Article
    • MEDLINE
    • CrossRef
25. Castaldi JL, Bertin MS, Giminez F, et al. Periodontal disease: is it a risk factor for premature labor, low birth weight or preeclampsia?. Rev Panam Salud Publica. 2006;19:253–258
    • View In Article
    • MEDLINE
    • CrossRef
26. American College of Obstetrics and Gynecology. Practice bulletin no. 33 (Diagnosis and management of preeclampsia and eclampsia). 2002;January
    • View In Article
27. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol. 1996;87:163–168
    • View In Article
    • MEDLINE
    • CrossRef
28. Michalowicz BS, Hodges JS, DiAngelis AJ, et al. OPT Study (Treatment of periodontal disease and the risk of preterm birth). N Engl J Med. 2006;355:1885–1894
    • View In Article
    • CrossRef
29. Manau C, Echeverria A, Agueda A, Guerrero A, Echeverria JJ. Periodontal disease definition may determine the association between periodontitis and pregnancy outcomes. J Clin Periodontol. 2008;35:385–397
    • View In Article
    • CrossRef