Study Aims:
The primary aim of the study is to determine the effect of a package of oral health interventions in pregnancy on the incidence of small vulnerable newborns (SVN).
* Hypothesis 1a: The oral health intervention will reduce overall incidence of SVNs at birth and individual outcomes of preterm birth, small-for-gestational age (SGA), and low birth weight (LBW).
* Hypothesis 1b: The intervention will have a larger effect on high-risk groups, including women who are younger/older, undernourished (low body mass index, low dietary diversity), or have more severe periodontal disease at enrollment.
* Hypothesis 1c: Adherence to the intervention will modify the effect of the intervention on the primary and secondary outcomes in a dose-response relationship.
Aim 2 will explore relationships between subgingival inflammation, systemic inflammation, and incidence of SVNs.
* Hypothesis 2a: The oral health intervention will reduce levels of subgingival inflammatory mediators between early and late pregnancy in the intervention group (IL-1β, IL-6, IL-8, TNF-α, etc.) relative to changes in the control group.
* Hypothesis 2b: Lower levels of subgingival inflammatory mediators will be associated with reduced risk for SVNs, shorter gestation, and low infant weight-for-age and weight at birth.
* Hypothesis 2c: Systemic inflammatory mediators will mediate the effect of subgingival inflammation levels on SVNs and adverse pregnancy outcomes.
Study Design:
This study is a community-based, unmasked, household-randomized controlled trial to assess the efficacy of a package of oral health interventions on the population-level incidence of SVN types in rural Sarlahi District, Nepal. The trial will enroll 2,280 pregnant women.
In a sub-study (n=200; 100 per intervention group), the investigators will collect venous blood and gingival crevicular fluid (GCF; 4 sites) in early and late pregnancy to explore relationships between subgingival inflammation, systemic inflammation, and SVNs and individual adverse pregnancy outcomes.
A sub-group of participants in the intervention and control arm will receive a periodontal exam in early (\<14 weeks) and late pregnancy (\~32-34 weeks).
Study Site and Participants:
The study will be conducted at the Nepal Nutrition and Health Intervention Project, Sarlahi (NNIPS) field research site in a rural community in the Terai region of Nepal. The Terai region (low lying plains) of Nepal has high rates of maternal undernutrition, neonatal mortality, preterm birth, SGA, and LBW. This area is representative of rural, low-income settings in South Asia. In this community, neonatal mortality is high (32 deaths per 1,000 live births).
A census of households in a defined geographic area of Sarlahi District covering areas of five municipalities (Haripur, Kabilasi, Chandranagar, Chakkargattha, and Ishwarpur), will be conducted to identify women of reproductive age (15-35 years at the time of the census). Consenting women will be followed via home visits every five weeks to assess menstrual period status since the previous visit. Women who report no menstrual period in the prior five weeks will be offered a pregnancy test. Those who test positive and are estimated to be \<12 weeks' gestation based on the date of last menstrual period (LMP) at the time of pregnancy identification will be invited to participate in the trial and enrolled as soon as possible.
Randomization:
Randomization will be done at the household level for this study. Once the first pregnant woman from a given household is identified and consents to be in the trial, that household will be allocated to either the intervention or control group. Any other woman from this household who subsequently become(s) pregnant within the pregnancy identification and enrollment period, will be assigned to the same group. Participants in the biospecimen substudy will be selected systematically towards the end of the trial enrollment.
Field Procedures and Data Collection:
Household census: A household census will be conducted by Field Interviewers (FIs). FIs will map all households and enumerate eligible married women 15-35 years in the study area using Google Earth satellite images and GPS coordinates. Next, we will visit each house to request consent from the head of household to identify if there are married women ages 15-35 living in the household. The investigators will consent women identified in this manner to participate in pregnancy surveillance. Women identified as living in a household who were not available during the household census visit will be added to a study roster, and an attempt to meet and consent the participant for pregnancy surveillance will be made during the first pregnancy surveillance round.
Pregnancy surveillance visits: Incident pregnancies will be identified through household visits every 5 weeks by a Community Health Research Worker (CHRW), a local woman from that village employed by NNIPS, to assess participant's menstrual status in the prior 30-day period and offer a urine-based pregnancy test.
Enrollment visit: FI will conduct an enrollment visit after being notified about a new pregnancy by CHRWs. At this visit, the FIs will conduct informed consent, administer questionnaires to collect data on demographics, socioeconomics, education/literacy, pregnancy and medical history, ANC care seeking, nutrition and diet diversity, household food security, tobacco and alcohol use, maternal depression; oral hygiene behaviors, dental health care seeking; and conduct anthropometric measurements, including maternal height (portable stadiometer), weight (MicroLife WS50 scale), mid-upper arm circumference (MUAC) (non-stretch insertion tape), blood pressure, and pulse.
Ultrasound visit: Shortly after pregnancy ascertainment and enrollment consent, Auxiliary Nurse Midwives (ANMs) will conduct the ultrasound gestational age estimation visit at participant homes. ANMs are trained to use a portable machine and follow the INTERGROWTH 21st methodology for early and late fetal gestational age assessment, including measurement of crown-rump length (CRL) (\<14 weeks) and biparietal diameter (outer-to-outer), head circumference, and femur length (≥14 weeks). The investigators will attempt to obtain all ultrasound measures \<14 weeks.
Early pregnancy periodontal examination visit: Half of participants in the intervention and control groups will receive this visit. The periodontal examiners will set up the equipment for the examination (dental chair, exam light, exam tools, sample collection kit, cool box, etc.) at the household. This will be a full-mouth periodontal examination, including measurement of probing pocket depth at six sites per tooth (disto-, mid-, and mesial- aspects of buccal and lingual surfaces) and the cemento-enamel junction to gingival margin distance (CEJ-GM) on two sites per tooth (mid- buccal and lingual aspects), excluding third molars. After probing each quadrant, data collectors will assess presence or absence of bleeding on probing (BOP) for buccal and lingual surfaces of each tooth. PPD values will be recorded in millimeters from 1 to 10, rounded to the next higher whole number. CEJ-GM distances will be recorded similarly, with values of 0 to 10 mm. If the free gingiva is coronal to the CEJ, the CEJ-GM measurement will be recorded as 0. Clinical attachment loss (CAL) will be calculated by summing the PD and CEJ-GM distance at the mid-site.
Pregnancy follow-up visits: CHRWs or FIs will visit women in the intervention group twice monthly during pregnancy to assess compliance and deliver additional antiseptic rinse, toothpaste, and toothbrush in the intervention group and collect data on maternal morbidities. At these visits, the data collectors will ask to observe the oral rinse use and brushing behaviors at periodic intervals. In instances where participants have questions or need to improve behaviors, the data collectors will provide a refresher on the instructions.
Pregnancy follow-up visits: CHRWs or FIs will visit women in the control group once monthly during pregnancy to collect data on maternal morbidities and ANC visits. At the first visit, data collectors will provide women with a manual toothbrush, toothpaste, and basic educational information about oral hygiene.
Late pregnancy visit: FIs will conduct a late pregnancy visit between 32- and 34-weeks' gestation to collect information on maternal anthropometry (same measures as enrollment), maternal morbidities, and delivery plans of the woman.
Late pregnancy periodontal examination visit: Half of participants in the intervention and control groups will receive this visit to determine if the intervention improved gingival health and reduced inflammation. Periodontal examiners will conduct the same exam as in early pregnancy.
Birth notification: At the start of the third trimester, data collectors will provide pregnant participants with a card that lists the cell phone number of the mobile birth team and instructions to notify the NNIPS staff as soon as the delivery occurs. CHRWs closely monitor each participant towards the end of pregnancy to ensure timely pregnancy outcome assessment.
Birth assessment visit: A specialized mobile team of FIs visit the household or hospital as soon as possible and within \<72 hours of delivery to conduct the birth interview, to collect information about the birth and delivery process, early breastfeeding and feeding behaviors, and initial newborn care practices. If the team is unable to visit the woman within 72 hours, the team member will still visit her when possible to collect information about the birth and early postpartum period and obtain anthropometric measures. Maternal and infant weight (g), length (cm), and head circumference (cm), are measured. Weight is measured with digital scales (Tanita BD-585), length with Shorr-type boards, and MUAC with non-stretch insertion tape. Non-digital measurements are taken in triplicate, and the median is used for analysis. The investigators will try to mask assessors to participant's allocation to the extent possible.
Sample Size:
This two-arm household-randomized controlled trial will require 2,280 pregnancies to detect a relative reduction of 15% in the incidence of SVN births between groups. To achieve this sample size, the investigators will enroll women over a 24-36-month period. Based upon data from previous NNIPS trials, the investigators assume a prevalence of SVN birth of 47% in the control group, 18% fetal loss rate and 7% loss to follow-up in pregnancy, and standard type I (5%) and type II (20%) errors.
Data analysis:
The investigators primary analysis for Aim 1 will utilize an intention-to-treat approach to determine the effect of the intervention package on the proportion of SVN births in each group. The investigators will construct the investigators primary and secondary outcome variables using first trimester obstetric ultrasound for gestational age and weights collected within \<72 hours for birth weight. For women without a first trimester ultrasound examination, which will not be common, the investigators will use either 2nd trimester ultrasound examination or LMP data for this purpose. For women/infant pairs reached for the anthropometry delivery visit after 72 hours and up to 10 days, the investigators will impute the missing early birth weight using an empirical Bayes model of early neonatal weight change developed by the investigators using data from a previous longitudinal study; this algorithm relies on other variables, including maternal education, age and parity, single or multiple pregnancy, infant sex, gestational age and neonatal survival status. the investigators will calculate the proportion of births, overall and by intervention group, that are SVN (any combination of preterm (\<37 weeks' gestation), SGA (\<10 centile of the INTERGROWTH-21st reference standard by gestational age and sex), and/or LBW (\<2,500g)) as well as the proportion of births affected by each individual secondary outcome, preterm, SGA, or LBW. The investigators will create histograms and line graphs to depict the distributions of continuous measures of infant gestational age, weight, and size-for-gestational age at birth overall and by intervention group.
To test H1a, The investigators will use a mixed effects log binomial regression model with an indicator variable for intervention group to estimate the relative risk and 95% confidence intervals of the primary outcome of SVN birth, accounting for clustering by the trial's design (if necessary), and adjusting for participant factors imbalanced across groups, as necessary. The investigators will apply a similar analytical approach in three additional models for each secondary outcome of preterm birth, SGA birth, or LBW birth. The investgigators will explore the intervention effect on the four-category definition of SVN types, T+SGA, PT+AGA, and PT+SGA, using mixed effects multinomial regression modeling (T+AGA (reference)).
For Aim 2, the investigatos will isolate the GCF inflammatory mediator specimen from the PerioCol strips by eluting to a standard volume (diluent 43 (Mesoscale Discovery)) and centrifuging according to established protocols. The investigators will use custom 96-well plate multiplex immunoassays (U-PLEX, Meso Scale Discovery) to quantify inflammatory mediators, including at least IL-1β, IL-4, IL-6, IL-8, IL-10, IL-17, IL-12p70, tumor necrosis factor- α (TNF-α), macrophage inflammatory protein-1α (MIP-1α), and C- reactive protein (CRP), measured in duplicate according to the manufacturer's standards. These mediators were selected for implication in potential causal pathways between with periodontal disease and adverse pregnancy outcomes, and this list will be finalized prior through a literature review of sources current at the start of the study. The investigators will determine from the assays the total amount of each inflammatory mediator and calculate the concentration using the volume of GCF obtained from the Periotron 8010 reading. The investigators will assess the normality of the distribution of inflammatory mediators using the Shapiro-Wilk Test and apply transformations to the data, as appropriate. the investigators will calculate descriptive statistics for inflammatory mediators, including mean ± SD, median, interquartile range, and depicted as scatter plots. The investigators will compare baseline inflammatory status across intervention groups to assess for imbalance. Analyses will be adjusted confounders as necessary and for multiple comparisons using the Benjamini-Hochberg procedure. Investigations for each inflammatory mediator will be conducted separately as follows.
To test H2a, the investigators will compare the difference in the change in GCF inflammatory mediator levels between early and late pregnancy in the intervention group compared to the control group. Separate mixed effects linear regression models will be run for each inflammatory mediator to estimate the median difference between groups and associated 95% confidence intervals, accounting for the trial's cluster design and adjusting for multiple comparisons. The investigators will also compare the difference in median concentration of inflammatory mediators at late pregnancy (independent of change from early pregnancy) between the intervention and control group using a similar regression approach.
