Risk of preterm birth associated with maternal gingival inflammation and oral hygiene behaviours in rural Nepal: a community-based, prospective cohort study

Objectives Observational studies have identified associations between periodontitis and adverse pregnancy outcomes, but randomised controlled trials evaluating the efficacy of periodontal therapy have yielded inconsistent results. Few studies have explored relationships between gingival inflammation and these outcomes or been conducted in rural, low-income communities, where confounding risk factors differ from other settings. Methods We conducted a community-based, prospective cohort study with the aim of estimating associations between the extent of gingival inflammation in pregnant women and incidence of preterm birth in rural Nepal. Our primary exposure was gingival inflammation, defined as bleeding on probing (BOP) ≥10%, stratified by BOP <30% and BOP ≥30%. A secondary exposure, mild periodontitis, was defined as ≥2 interproximal sites with probing depth (PD) ≥4 mm (different teeth) or one site with PD ≥5 mm. Our primary outcome was preterm birth (<37 weeks gestation). We used Poisson regression to model this relationship, adjusting for potential confounders. Results Of 1394 participants, 554 (39.7%) had gingival inflammation, 54 (3.9%) mild periodontitis and 197 (14.1%) delivered preterm. In the adjusted regression model, increasing extent of gingival inflammation was associated with a non-significant increase in risk of preterm birth (BOP ≥30% vs no BOP: adjusted relative risk (aRR) 1.37, 95% CI: 0.81 to 2.32). A secondary analysis, stratifying participants by when in pregnancy their oral health status was assessed, showed an association between gingival inflammation and preterm birth among women examined in their first trimester (BOP ≥30% vs no BOP: aRR 2.57, 95% CI: 1.11 to 5.95), but not later in pregnancy (BOP ≥30% vs no BOP: aRR 1.05, 95% CI: 0.52 to 2.11). Conclusions Gingival inflammation in women examined early in pregnancy and poor oral hygiene behaviours were risk factors for preterm birth. Future studies should evaluate community-based oral health interventions that specifically target gingival inflammation, delivered early in or before pregnancy, on preterm birth. Trial registration number Nepal Oil Massage Study, NCT01177111.

I, the Submitting Author has the right to grant and does grant on behalf of all authors of the Work (as defined in the below author licence), an exclusive licence and/or a non-exclusive licence for contributions from authors who are: i) UK Crown employees; ii) where BMJ has agreed a CC-BY licence shall apply, and/or iii) in accordance with the terms applicable for US Federal Government officers or employees acting as part of their official duties; on a worldwide, perpetual, irrevocable, royalty-free basis to BMJ Publishing Group Ltd ("BMJ") its licensees and where the relevant Journal is co-owned by BMJ to the co-owners of the Journal, to publish the Work in this journal and any other BMJ products and to exploit all rights, as set out in our licence.
The Submitting Author accepts and understands that any supply made under these terms is made by BMJ to the Submitting Author unless you are acting as an employee on behalf of your employer or a postgraduate student of an affiliated institution which is paying any applicable article publishing charge ("APC") for Open Access articles. Where the Submitting Author wishes to make the Work available on an Open Access basis (and intends to pay the relevant APC), the terms of reuse of such Open Access shall be governed by a Creative Commons licence -details of these licences and which Creative Commons licence will apply to this Work are set out in our licence referred to above.
Other than as permitted in any relevant BMJ Author's Self Archiving Policies, I confirm this Work has not been accepted for publication elsewhere, is not being considered for publication elsewhere and does not duplicate material already published. I confirm all authors consent to publication of this Work and authorise the granting of this licence. Annually, 2.5 million babies die prior to 28 days of life, and preterm birth is the leading cause of these deaths. [1] Preterm newborns that survive are at substantial risk of mortality from other causes, long-term disabilities such as neurological and developmental impairments, and noncommunicable diseases. [2] In low-and middle-income countries (LMICs), where the majority of preterm births occur, therapeutic interventions are often unavailable and difficult to scale up, especially in communities where many mothers deliver at home or in primary facilities without skilled care (e.g. South Asia). [3] Periodontal disease includes several inflammatory conditions, typically initiated by oral bacteria, beginning with reversible accumulation of plaque and inflammation of gingival tissue (gingivitis) and progressing to irreversible breakdown of the supportive tissues of the teeth and tooth loss (periodontitis). [4] Onset of new and worsening of existing gingival inflammation during pregnancy are normal and well documented, peaking in the second or third trimester. [5,6] Major physiological and hormonal changes occur during pregnancy with wide ranging effects on the body, including increased permeability of gingival capillaries, altered immune system activity, and shifts in composition of the sub-gingival microbiome, including proliferation of aggressive bacteria associated with periodontitis, such as Porphyromonas gingivalis. [7,8] Periodontal disease in pregnant women has been associated with preterm birth and other adverse pregnancy outcomes. [9][10][11] Randomized controlled trials (RCTs) evaluating the impact of periodontal therapy during pregnancy on adverse pregnancy outcomes have produced inconsistent results. [9,[12][13][14][15][16][17][18] One meta-analysis found a significant effect of periodontal therapy among women at high risk of preterm birth. [19] Although the mechanisms underlying the observed association are unclear, hypotheses include hematogenic translocation of periodontal pathogens or their byproducts to the fetal-placental unit, or action of inflammatory mediators in the periodontium on levels of systemic inflammation. [20] Alternatively, the observed relationship could be due to a common confounding factor, such as a genetic hyper-inflammatory phenotype, responsible for both increased risk for periodontal disease and adverse pregnancy outcomes. [21] Few studies have evaluated whether the extent or severity of gingival inflammation are associated with risk of preterm birth or other adverse pregnancy outcomes. Further, studies of this association have been nearly universally facility-based, whether in high-or low-income settings. Understanding this relationship from a population-based perspective in low resource communities can offer certain benefits from an epidemiological perspective. Many populations in LMICs have lower prevalence of important confounding factors of this relationship, such as smoking, alcohol use, and chronic diseases (e.g. hypertension or diabetes). Community-based studies can avoid selection bias associated with hospital-based studies, particularly in populations where home delivery remains common, as is the case in much of South Asia. Given these potential benefits, we conducted a community-based, prospective cohort study to estimate the association between gingival inflammation and preterm birth among women in a rural community in the terai (plains) region of Nepal.

METHODS
We enrolled pregnant women <26 weeks gestation in a community-based, prospective cohort study of maternal gingival inflammation and adverse pregnancy outcomes across nine village development committees (VDCs) in Sarlahi District, Nepal, between January and November 2016. Participants in this study were identified and determined eligible using the infrastructure of a large community-based randomized trial of topical applications for newborn massage, the Nepal Oil Massage Study (NOMS) (NCT01177111), which was actively enrolling a populationbased sample of pregnant women in this study area.

Study visits
Study visits were conducted in participant homes because of the wide dispersion of households across this rural community and the impracticality of bringing participants to a central location.
Data on participant demographics, vital signs and morbidities during pregnancy, oral hygiene practices, care-seeking, and knowledge, and other characteristics were collected through a series of questionnaires administered over several visits during the course of pregnancy. Data collection teams were notified of the birth outcome by a locally resident study staff member, and the date of birth and other data concerning the mother and newborn were collected as soon as possible after delivery.
Oral health examinations were performed by five auxiliary nurse-midwives who were trained as   8 We defined secondary exposure definitions for mild and moderate periodontitis based upon a modified version of Centers for Disease Control and AAP updated 2012 case definitions for population-based surveillance of periodontitis. [27,29] Mild periodontitis was defined as ≥2 interproximal sites with PD≥4 mm (not on the same tooth) or one site with PD≥5 mm. Moderate periodontitis was defined as ≥2 interproximal sites with PD≥5 mm (not on the same tooth). CAL was not included in these definitions as we did not collect this measure at interproximal sites to shorten the visit time; however, clinical attachment level and probing depth are widely considered equivalent measures of periodontitis in younger adults. [30] Gestational age was calculated using the last menstrual period (LMP) method as recalled by the mother at 5-weekly pregnancy surveillance home visits. Our 5-weekly pregnancy surveillance approach has several benefits, including short recall (≤5 weeks) and pregnancy testing, as compared to traditional LMP approaches, which often rely on LMP recall later in pregnancy or at time of delivery. Preterm birth was defined as a live birth or stillbirth <37 completed weeks of gestation at the time of delivery.

Statistical analysis
Bivariate analyses between participant characteristics and the outcome, preterm birth, were evaluated using t-tests and Poisson regression for continuous and binary/categorical variables, respectively. We calculated unadjusted and adjusted relative risks (RR & aRR) of preterm birth and associated 95% confidence internals (CI) using Poisson regression with robust variance, which was used due to occurrence of convergence issues with other regression methods. Given our sample size (n=1,394) and assuming a power of 80%, type I error of 0.05, and our pre-study estimates of population prevalence of preterm birth (17%) and gingival inflammation (40%), we estimated that we could detect a relative risk of preterm birth of roughly >1.5.
Multivariable models were constructed by sequentially adding groups of covariates, beginning with maternal characteristics, oral hygiene behaviors, and socioeconomic factors. Covariates associated with preterm birth at the p<0.10 level in bivariate analyses were considered in these regression models. Additional variables, known through previous studies to be confounders of the periodontal disease and preterm birth relationship, were also included in regression models (including age, ethnicity, body mass index (BMI), primiparity, multiple births, and socioeconomic variables (i.e., literacy, education, and indicators of household wealth).
In an effort to remove a possibly attenuating effect of including cases of transient pregnancyinduced gingival inflammation, we examined the relationship stratified by timing of oral exam.
Women were categorized into two trimester groups (<13 weeks and ≥13 to <26 weeks gestation).
Similar to the primary analysis, for each trimester group we conducted adjusted relative risks (RR & aRR) of preterm birth and associated 95% confidence internals (CI) using Poisson regression with robust variance.

Participant characteristics
Between January 11, 2016, and November 26, 2016, among 2,821 pregnancies in the study area, 2,291 (81.2%) participants were identified by the parent trial as eligible for enrollment (<26 weeks gestation) (Figure 1). Among eligible participants, 1,478 (64.5%) were enrolled and six refused participation (0.3%). Another 807 (35.2%) were not visited for the purposes of the study because the eligible participants in the study area initially exceeded the capacity of the oral health workers to enroll women before their gestational age exceeded the exclusion criterion of <26 weeks. For logistical reasons, enrollment of women earlier in pregnancy was prioritized at the start of study enrollment, while enrollment of women later in pregnancy was prioritized towards the end of study enrollment. However, among eligible women, the included and excluded participant groups did not differ by age, education, or socio-economic factors.
Birth outcomes for the 1,474 women followed until the end of pregnancy were recorded as 1,345 single live births, 14 twin live births, 33 single stillbirths, 1 twin stillbirth, 1 set of twins where 1 was live born and 1 stillborn, 74 miscarriages, and 6 abortions. As our primary outcome was preterm birth, pregnancy was selected as the unit of analysis (and, therefore, twins were counted only once). There were 1,394 total pregnancies for analysis, of which 197 (14.1%) were preterm.
Few observational or interventional studies have considered gingival inflammation as the primary exposure definition when investigating the relationship between periodontal disease and preterm birth. A cohort study by Kruse et al. (2008) found an association between gingivitis and high risk of preterm birth among women without periodontitis in a hospital setting in to women examined earlier in pregnancy, this could have resulted in non-differential misclassification of our binary outcome, preterm birth, which would tend to attenuate observation of a true relationship among women examined in the second trimester.

CONCLUSION
Our study identified gingival inflammation as an independent risk factor for preterm birth among women examined early in pregnancy. Women non-compliant with oral hygiene care recommendations were at substantially increased risk of preterm birth, highlighting the importance of adherence to proper home care in pregnant women and women expecting to become pregnant. Future studies should evaluate the effectiveness of community-based oral health interventions that specifically target gingival inflammation-delivered to women early in or prior to pregnancy -on the incidence of preterm birth and other adverse pregnancy outcomes in low-income countries with high risk for these outcomes. Dr. Erchick conceptualized and designed the study, developed field implementation protocols, led data collection in the field, conducted the analysis, and wrote the manuscript.
Dr. Khatry conceptualized and designed the study, oversaw field implementation and ensured quality data collection, and provided comments on the manuscript.
Dr. Agrawal conceptualized and designed the study, trained and oversaw the data collectors, and provided comments on the manuscript.
Dr. Katz conceptualized and designed the study, ensured quality data collection, advised on analytic approach, and provided comments on the manuscript.
Mr. LeClerq conceptualized and designed the study, supported overall implementation in the field, and provided comments on the manuscript.
Mr. Rai contributed to field implementation, ensured quality control of field procedures, and provided comments on the manuscript.
Dr. Reynolds conceptualized and designed the study, contributed to the analysis and interpretation of the results, and provided comments on the manuscript.
Dr. Mullany conceptualized and designed the study, obtained funding for the study, oversaw implementation of data collection, obtained ethical approvals, advised on analytic approach, and provided edits and comments on the manuscript.
All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.   Conclusions: Gingival inflammation in women examined early in pregnancy and poor oral hygiene behaviors were risk factors for preterm birth. Future studies should evaluate communitybased oral health interventions that specifically target gingival inflammation, delivered early in or before pregnancy, on preterm birth.

Strengths and Limitations of this Study
 This study collected data through a large, prospective, community-based cohort in a rural, low-resource setting.
 The study population had low prevalence of some established confounders of the relationship of interest, which are common in other populations.
 Some clinical periodontal measures were not collected due to visit time constraints, for example, recession on interproximal sites or gingival index score.
 Preterm birth classification was based upon maternal self-report of last menstrual period (LMP) instead of the gold standard, ultrasound examination. Periodontal disease includes several inflammatory conditions, typically initiated by oral bacteria, beginning with reversible accumulation of plaque and inflammation of gingival tissue (gingivitis) and progressing to irreversible breakdown of the supportive tissues of the teeth and tooth loss (periodontitis). [4] Onset of new and worsening of existing gingival inflammation during pregnancy are normal and well documented, peaking in the second or third trimester. [5,6] Major physiological and hormonal changes occur during pregnancy with wide ranging effects on the body, including increased permeability of gingival capillaries, altered immune system activity, and shifts in composition of the sub-gingival microbiome, including proliferation of aggressive bacteria associated with periodontitis, such as Porphyromonas gingivalis. [7,8] Periodontal disease in pregnant women has been associated with preterm birth and other adverse pregnancy outcomes. [9][10][11] Randomized controlled trials (RCTs) evaluating the impact of periodontal therapy during pregnancy on adverse pregnancy outcomes have produced inconsistent results. [9,[12][13][14][15][16][17][18] One meta-analysis found a significant effect of periodontal therapy

Study visits
Study visits were conducted in participant homes because of the wide dispersion of households across this rural community and the impracticality of bringing participants to a central location.
Data on participant demographics, vital signs and morbidities during pregnancy, oral hygiene practices, care-seeking, and knowledge, and other characteristics were collected through a series of questionnaires administered over several visits during the course of pregnancy. Data collection teams were notified of the birth outcome by a locally resident study staff member, and the date of birth and other data concerning the mother and newborn were collected as soon as possible after delivery.
Oral health examinations were performed by five auxiliary nurse midwives who were trained for the purpose of this study. Their training lasted 3-4 weeks and included classroom instruction and practice of periodontal techniques conducted by an experienced dentist (NKA) at the Department

Periodontal measurements
Auxiliary nurse midwives used portable dental equipment to conduct a full mouth examination.
Periodontal measurements were made using a color Williams probe (Hu-Friedy, Chicago, IL, USA). PD was measured on six sites per tooth (disto-, mid-, and mesial-aspects of buccal and lingual surfaces) and CEJ-GM distance on two sites per tooth (mid-buccal and lingual aspects), excluding third molars. Presence or absence of BOP was recorded for any buccal or lingual probing site of each tooth. PD values were recorded in millimeters rounded to the next higher whole number. CEJ-GM distances were recorded similarly. If the free gingiva was coronal to the CEJ, the CEJ-GM measurement was recorded as 0. Clinical attachment loss (CAL) was calculated by summing PD and CEJ-GM distances.

Outcome/exposure definitions
Our primary exposure definition, gingival inflammation, was defined according to a classification scheme developed by the American Academy of Periodontology (AAP) and European Federation of Periodontology to update the 1999 classification of periodontal diseases and conditions. [23] Clinical health was defined as all sites PD ≤3 mm and BOP <10%. [24,25] Gingival inflammation was defined as BOP ≥10%, stratified as localized gingival inflammation (BOP 10%-30%) and generalized gingival inflammation (BOP ≥30%). [24][25][26][27][28][29]  Gestational age was calculated using the last menstrual period (LMP) method as recalled by the mother at 5-weekly pregnancy surveillance home visits. Our 5-weekly pregnancy surveillance approach has several benefits, including short recall (≤5 weeks) and pregnancy testing, as compared to traditional LMP approaches, which often rely on LMP recall later in pregnancy or at time of delivery. Preterm birth was defined as a live birth or stillbirth <37 completed weeks of gestation at the time of delivery.

Statistical analysis
Bivariate analyses between participant characteristics and the outcome, preterm birth, were evaluated using t-tests and Poisson regression for continuous and binary/categorical variables, respectively. We calculated unadjusted and adjusted relative risks (RR & aRR) of preterm birth and associated 95% confidence internals (CI) using Poisson regression with robust variance, which was used due to occurrence of convergence issues with other regression methods. Given In an effort to remove a possibly attenuating effect of including cases of transient pregnancyinduced gingival inflammation, we examined the relationship stratified by timing of oral exam.
Women were categorized into two trimester groups (<13 weeks and ≥13 to <26 weeks gestation).
Similar to the primary analysis, for each trimester group we conducted adjusted relative risks   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 F o r p e e r r e v i e w o n l y   Our secondary exposure definition for mild periodontitis was not associated with preterm in a binary comparison or the crude or adjusted regression models for our primary or secondary analyses. These analyses were not conducted for moderate periodontitis given the low prevalence of the condition in this study population.

DISCUSSION
In our community-based, prospective cohort study, gingival inflammation was an independent risk factor for preterm birth among women examined during their first, but not second, trimester, after adjusting for potential confounding risk factors. Our finding suggests that gingival infection and inflammatory burden early in pregnancy, potentially originating prior to conception, could be responsible for the observed relationship, which becomes obscured in the second trimester by the presence of women with initially healthy periodontal conditions who develop pregnancyinduced gingival inflammation. This result is consistent with studies that have reported that active, and especially progressive, periodontal infection is most harmful early in pregnancy. [13] This also supports the hypothesis, proposed to explain the mixed results of RCTs of this association, that intervening on women with periodontal disease later in pregnancy, or even after conception, may be too late to affect the proposed causal pathway. [33] Most previous studies have considered the relationship between varying definitions and severities of periodontitis and preterm birth. A meta-analysis of seven cohort studies estimated a pooled relative risk of preterm of 1 Unlike the majority of studies of this association, particularly in high-income countries, the prevalence of important confounders, such as smoking, alcohol use, and morbidities associated with chronic disease, were nearly absent in our study population. An association between periodontitis and preterm birth may have been seen in our population had there been a higher prevalence and severity of periodontitis, as would be expected in an older population.  to women examined earlier in pregnancy, this could have resulted in non-differential misclassification of our binary outcome, preterm birth, which would tend to attenuate observation of a true relationship among women examined in the second trimester.

CONCLUSION
Our study identified gingival inflammation as an independent risk factor for preterm birth among women examined early in pregnancy. Women non-compliant with oral hygiene care recommendations were at substantially increased risk of preterm birth. These findings highlight the importance of adherence to proper home oral self-care in pregnant women and women expecting to become pregnant. Oral health policies and education programs should include tailored approaches to encourage improved self-care practices in this important population.
Future studies should evaluate the effectiveness of community-based oral health interventions that specifically target gingival inflammation -delivered to women early in or prior to pregnancy -on the incidence of preterm birth and other adverse pregnancy outcomes in low-income countries with high risk for these outcomes. Patients or public involvement: Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

Authors' contribution:
Dr. Erchick conceptualized and designed the study, developed field implementation protocols, led data collection in the field, conducted the analysis, and wrote the manuscript.
Dr. Khatry conceptualized and designed the study, oversaw field implementation and ensured quality data collection, and provided comments on the manuscript.
Dr. Agrawal conceptualized and designed the study, trained and oversaw the data collectors, and provided comments on the manuscript.
Dr. Katz conceptualized and designed the study, ensured quality data collection, advised on analytic approach, and provided comments on the manuscript.
Mr. LeClerq conceptualized and designed the study, supported overall implementation in the field, and provided comments on the manuscript.
Mr. Rai contributed to field implementation, ensured quality control of field procedures, and provided comments on the manuscript.

Strengths and Limitations of this Study
 This study collected data through a large, community-based, prospective cohort in a rural, low-resource setting.
 The study population had low prevalence of some established confounders of the relationship of interest, which are common in other populations.
 Some clinical periodontal measures were not collected due to visit time constraints, for example, recession on interproximal sites or gingival index score.
 Preterm birth classification was based upon maternal self-report of last menstrual period (LMP) instead of the gold standard, ultrasound examination.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y INTRODUCTION Annually, 2.5 million babies die prior to 28 days of life, and preterm birth is the leading cause of these deaths. [1] Preterm newborns that survive are at substantial risk of mortality from other causes, long-term disabilities such as neurological and developmental impairments, and noncommunicable diseases. [2] In low-and middle-income countries (LMICs), where the majority of preterm births occur, therapeutic interventions are often unavailable and difficult to scale up, especially in communities where many mothers deliver at home or in primary facilities without skilled care (e.g., South Asia). [3] Periodontal disease includes several inflammatory conditions, typically initiated by oral bacteria, beginning with reversible accumulation of plaque and inflammation of gingival tissue (gingivitis) and progressing to irreversible breakdown of the supportive tissues of the teeth and tooth loss (periodontitis). [4] Onset of new and worsening of existing gingival inflammation during pregnancy are normal and well documented, peaking in the second or third trimester. [5,6] Major physiological and hormonal changes occur during pregnancy with wide ranging effects on the body, including increased permeability of gingival capillaries, altered immune system activity, and shifts in composition of the sub-gingival microbiome, including proliferation of aggressive bacteria associated with periodontitis, such as Porphyromonas gingivalis. [7,8] Periodontal disease in pregnant women has been associated with preterm birth and other adverse pregnancy outcomes. [9][10][11] Yet randomized controlled trials (RCTs) evaluating the impact of periodontal therapy during pregnancy on adverse pregnancy outcomes have produced inconsistent results. [9,[12][13][14][15][16][17][18] One meta-analysis found a significant effect of periodontal therapy  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   5 among women at high risk of preterm birth. [19] Although the mechanisms underlying this observed association are unclear, hypotheses include hematogenic translocation of periodontal pathogens or their byproducts to the fetal-placental unit, or action of inflammatory mediators in the periodontium on levels of systemic inflammation. [20] Alternatively, the observed relationship between periodontal disease in pregnant women and adverse pregnancy outcomes could be the result of unmeasured and uncontrolled confounding factors. Previously described confounders of this relationship, which are commonly controlled for in studies, include age, smoking, multiple birth, previous adverse pregnancy outcomes, ethnicity, and socioeconomic factors. [11] Some studies, however, have proposed the possibility that a genetic inflammatory phenotype could be responsible for increased risk of periodontal disease, or failed periodontal theory, and adverse pregnancy outcomes, particularly spontaneous preterm birth. [21] Few studies have evaluated whether gingival inflammation is associated with risk of preterm birth or other adverse pregnancy outcomes. Further, studies of this association have been nearly universally facility-based, whether in high-or low-income settings. Understanding this relationship from a population-based perspective in low resource communities can offer certain benefits from an epidemiological perspective. Many populations in LMICs have lower prevalence of important confounding factors of this relationship, such as smoking, alcohol use, and chronic diseases (e.g. hypertension or diabetes). Community-based studies can avoid selection bias associated with hospital-based studies, particularly in populations where home delivery remains common, as is the case in much of South Asia. Given these potential benefits,  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   6 we conducted a community-based, prospective cohort study to estimate the association between gingival inflammation and preterm birth among women in a rural community in the Terai (plains) region of Nepal.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59

Study visits
Study visits were conducted in participant homes because of the wide dispersion of households across this rural community and the impracticality of bringing participants to a central location.
Data on participant demographics, vital signs and morbidities during pregnancy, oral hygiene practices, care-seeking, and knowledge, and other characteristics were collected through a series of questionnaires administered over several visits during the course of pregnancy. Data collection teams were notified of the birth outcome by a locally resident study staff member, and the date of birth and other data concerning the mother and newborn were collected as soon as possible after delivery.
Periodontal measurements were made using a color Williams probe (Hu-Friedy, Chicago, IL, USA). PD was measured on six sites per tooth (disto-, mid-, and mesial-aspects of buccal and lingual surfaces) and CEJ-GM distance on two sites per tooth (mid-buccal and lingual aspects), excluding third molars. Presence or absence of BOP was recorded for any buccal or lingual probing site of each tooth. PD values were recorded in millimeters rounded to the next higher whole number. CEJ-GM distances were recorded similarly. If the free gingiva was coronal to the CEJ, the CEJ-GM measurement was recorded as 0. Clinical attachment loss (CAL) was calculated by summing PD and CEJ-GM distances.

Outcome/exposure definitions
Our primary exposure definition, gingival inflammation, was defined according to a classification scheme developed by the American Academy of Periodontology (AAP) and European Federation of Periodontology to update the 1999 classification of periodontal diseases and conditions. [23] Clinical health was defined as all sites PD ≤3 mm and BOP <10%. [24,25] Gingival inflammation was defined as BOP ≥10%, stratified as localized gingival inflammation (BOP 10%-30%) and generalized gingival inflammation (BOP ≥30%). [24][25][26][27][28][29] 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   9 We defined secondary exposure definitions for mild and moderate periodontitis based upon a modified version of Centers for Disease Control and AAP updated 2012 case definitions for population-based surveillance of periodontitis. [28,30] Mild periodontitis was defined as ≥2 interproximal sites with PD≥4 mm (not on the same tooth) or one site with PD≥5 mm. Moderate periodontitis was defined as ≥2 interproximal sites with PD≥5 mm (not on the same tooth). CAL was not included in these definitions as we did not collect this measure at interproximal sites to shorten the visit time; however, clinical attachment level and probing depth are widely considered equivalent measures of periodontitis in younger adults. [31] Gestational age was calculated using the last menstrual period (LMP) method as recalled by the mother at 5-weekly pregnancy surveillance home visits. Our 5-weekly pregnancy surveillance approach has several benefits, including short recall (≤5 weeks) and pregnancy testing, as compared to traditional LMP approaches, which often rely on LMP recall later in pregnancy or at time of delivery. Preterm birth was defined as a live birth or stillbirth <37 completed weeks of gestation at the time of delivery.

Statistical analysis
Bivariate analyses between participant characteristics and the outcome, preterm birth, were evaluated using t-tests and Poisson regression for continuous and binary/categorical variables, respectively. We calculated unadjusted and adjusted relative risks (RR & aRR) of preterm birth and associated 95% confidence internals (CI) using Poisson regression with robust variance, which was used due to occurrence of convergence issues with other regression methods. Given In an effort to remove a possibly attenuating effect of including cases of transient pregnancyinduced gingival inflammation, we examined the relationship stratified by timing of oral exam.
Women were categorized into two trimester groups (<13 weeks and ≥13 to <26 weeks gestation).

DISCUSSION
In our community-based, prospective cohort study, gingival inflammation was an independent risk factor for preterm birth among women examined during their first, but not second, trimester, after adjusting for potential confounding risk factors. Our finding suggests that gingival infection and inflammatory burden early in pregnancy, potentially originating prior to conception, could be responsible for the observed relationship, which becomes obscured in the second trimester by the presence of women with initially healthy periodontal conditions who develop pregnancyinduced gingival inflammation. This result is consistent with studies that have reported that active, and especially progressive, periodontal infection is most harmful early in pregnancy. [13] This also supports the hypothesis, proposed to explain the mixed results of RCTs of this association, that intervening on women with periodontal disease later in pregnancy, or even after conception, may be too late to affect the proposed causal pathway. [33] Most previous studies have considered the relationship between varying definitions and severities of periodontitis and preterm birth. A meta-analysis of seven cohort studies estimated a pooled relative risk of preterm of 1  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  Poor oral hygiene behaviors, including extended teeth cleaning time (≥5 mins), infrequent teeth cleaning (<7 times per week), and use of traditional dentifrice (i.e., sand, ash, oil, or gul) or datiwan, were also associated with increased risk of preterm birth in our analysis. Studies have shown that the mechanical manipulation involved in brushing, particularly frequent or forceful brushing or use of brushes with hard filaments, can lead to gingival abrasion, recession, and bacteremia. [42][43][44] Some have posited that bacteremia caused by mechanical manipulation of the gingiva involved in periodontal therapy may increase risk of adverse pregnancy outcomes, attenuating observation of any true effect in interventional studies of this association. [33] More research is needed to understand the extent to which poor oral hygiene behaviors in pregnant women, including common traditional practices, such as datiwan use, could -via local An important strength of our study was that we utilized a population-based sample, compared to previous observational studies of this relationship, which were primarily conducted in health facility settings, introducing risk of selection bias associated with the likelihood of women in the target population delivering in a particular facility. The potential for selection bias to influence the measure of association in a facility-based study is especially high in populations where a substantial proportion of women deliver at home, as is the case in this area of rural Nepal, where roughly half of women deliver in a facility. In the unpublished review conducted by our research team, we found one study of this association, by , that used a populationbased sample. [45] Conducted in a periurban area of Hyderabad, Pakistan, the authors reported significant associations of varying degrees between measures of periodontal disease and neonatal death, perinatal death, and stillbirth. [45] A limitation of this study was the collection of clinical recession measures from only the direct buccal and lingual surfaces. In the absence of these data, we likely underestimated the burden of periodontal disease among pregnant women in this study, potentially attenuating our measure of association. We were unable to control for some confounders of this relationship, including previous preterm birth, certain chronic diseases (e.g., diabetes), and we only used a proxy (self-   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   29 and logistical constraints of our study. Lastly, preterm birth was based upon maternal self-report of last menstrual period (LMP) instead of ultrasound examination. If our LMP estimates were less accurate among women examined later in pregnancy, due to longer maternal recall relative to women examined earlier in pregnancy, this could have resulted in non-differential misclassification of our binary outcome, preterm birth, which would tend to attenuate observation of a true relationship among women examined in the second trimester.

CONCLUSION
Our study identified gingival inflammation as an independent risk factor for preterm birth among women examined early in pregnancy. Women non-compliant with oral hygiene care recommendations were at substantially increased risk of preterm birth. These findings highlight the importance of adherence to proper home oral self-care in pregnant women and women expecting to become pregnant. Oral health policies and education programs should include tailored approaches to encourage improved self-care practices in this important population.
Future studies should evaluate the effectiveness of community-based oral health interventions that specifically target gingival inflammation -delivered to women early in or prior to pregnancy -on the incidence of preterm birth and other adverse pregnancy outcomes in low-income countries with high risk for these outcomes. Standard deviation (SD)

Authors' contribution:
Dr. Erchick conceptualized and designed the study, developed field implementation protocols, led data collection in the field, conducted the analysis, and wrote the manuscript.
Dr. Khatry conceptualized and designed the study, oversaw field implementation and ensured quality data collection, and provided comments on the manuscript.
Dr. Agrawal conceptualized and designed the study, trained and oversaw the data collectors, and provided comments on the manuscript.
Dr. Katz conceptualized and designed the study, ensured quality data collection, advised on analytic approach, and provided comments on the manuscript.
Mr. LeClerq conceptualized and designed the study, supported overall implementation in the field, and provided comments on the manuscript.
Mr. Rai contributed to field implementation, ensured quality control of field procedures, and provided comments on the manuscript.
Dr. Reynolds conceptualized and designed the study, contributed to the analysis and interpretation of the results, and provided comments on the manuscript. implementation of data collection, obtained ethical approvals, advised on analytic approach, and provided edits and comments on the manuscript.