A pre-conception clinical trial to reduce intergenerational obesity and diabetes risks: the NDPP-NextGen trial protocol

Abstract

Background.

Intrauterine exposure to maternal overweight/obesity or diabetes transmits risks to offspring, perpetuating a disease cycle across generations. Prenatal interventions to reduce maternal weight or dysglycemia have limited impact, while postpartum interventions can alter the intrauterine environment only if child-bearing continues. Efficacious preconception interventions are needed, especially for underserved populations, and with the potential to be scaled up sustainably. Research is also needed to assess intervention effects at conception, throughout pregnancy, and among offspring.

Methods.

This two-arm, parallel randomized clinical trial will include 360 biological females with overweight/obesity and moderate-to-high likelihood of pregnancy within 24 months. Participants will be randomized 1:1 to a yearlong pre-conception lifestyle intervention based on the National Diabetes Prevention Program (NDPP-NextGen) or usual care. Data collection will occur at enrollment (before conception), post-conception (<8 weeks gestation), late pregnancy (28–32 weeks gestation), and delivery (before discharge) for participants who become pregnant within 24 months of enrollment. Main outcomes are post-conception body mass index (<8 weeks gestation; primary outcome), post-conception fasting glucose (<8 weeks gestation; secondary outcome), and neonatal adiposity (<2 days post-birth). Additional clinical, behavioral, perinatal and offspring data will be collected, and biospecimens (blood, urine, stool, cord blood) will be banked for future ancillary studies.

Conclusion.

This clinical trial will evaluate an intervention model (NDPP-NextGen) with potential to improve maternal health among the >50% of US females with overweight/obesity or diabetes risks in pregnancy. If successful, it can be scaled among >1800 organizations delivering NDPP in the United States to benefit the health of future generations.

Introduction

Intrauterine exposure to maternal overweight/obesity and diabetes transmits risks to offspring, which perpetuates a disease cycle across generations [1, 2]. Over half of adult females in the United States (US) have overweight/obesity at conception and ~10% of pregnancies are affected by diabetes, exposing nearly 2 million infants each year [3–5]. When offspring of females with overweight, obesity, or hyperglycemia enter their own reproductive years with these conditions, the next generation is similarly exposed, triggering a self-perpetuating cycle [6]. Thus, effective intervention strategies to improve maternal weight and glycemia during the childbearing years are needed.

Prenatal interventions have been well-studied in >100 trials across the globe [1], but are considered “too little, too late” to meaningfully impact maternal-child health as they typically begin in the second trimester [7]. While prenatal interventions tend to reduce maternal gestational weight gain (GWG) by ~1 kg and gestational diabetes (GDM) incidence by 20–30%, their impact on offspring is very limited [1]. Postpartum interventions (in the first year after delivery) and inter-conception programs (between pregnancies) often show 1–3 kg greater postpartum weight loss with intervention [8, 9], but with varying reduction in risk for developing type 2 diabetes after GDM (0–58%) [10]. Moreover, postpartum interventions can alter intrauterine exposures only if participants go on to have more children. Pre-conception interventions (before an anticipated pregnancy) show initial weight loss of 3–9 kg and improved glycemia, but weight is often regained before conception (i.e., ~1–2 years after intervention) [11] and offspring birthweight is largely unchanged [11, 12]. Greater preconception weight loss (16–28 kg) through bariatric surgery has improved GDM and neonatal size (birthweight, large-for-gestational age) [13, 14], but entails high risks, costs, and has restrictive eligibility criteria [15]. Further research on lifestyle interventions in the preconception period is needed to impact population health.

Several knowledge and implementation gaps need to be addressed before implementing pre-conception interventions to scale. First, most prior studies tested new lifestyle interventions that are unlikely to be scaled up, even if efficacious, given the difficulty of adopting new clinical practices [16]. Adapting interventions that have already been scaled has more potential for population impact. Second, targeting females likely to become pregnant soon can facilitate intervention around the key period of conception to early pregnancy [17], yet sophisticated approaches to identify this population are lacking. More thorough screening protocols are needed, particularly in more diverse and under-resourced populations, among whom 60% of pregnancies are unplanned [18]. Third, studies must carefully assess maternal weight at conception, given regain is typical after intervention [17, 19]. Yet many prior studies lack rigorous measures of weight and glycemia at conception. Fourth, offspring outcomes are often understudied, and rarely include future obesity markers beyond birthweight. Alternatives like neonatal body composition should be considered, which is more sensitive to intrauterine exposures than birthweight [20, 21] and distinguishes between fat and lean mass, which have different implications for metabolic health [22].

We describe the protocol for a randomized clinical trial of a pre-conception lifestyle intervention designed to address these gaps. The trial will enroll females with moderate-to-high likelihood of pregnancy within 2 years, recruited from a low-income population with racial and ethnic diversity. Novel features of this trial include leveraging an existing lifestyle modification program (the National Diabetes Prevention Program; NDPP) to facilitate scalability, a new approach to screen for likelihood of pregnancy, collecting weight at conception via cellular-enabled home scales, and measuring neonatal adiposity. We hypothesize that the intervention will reduce post-conception BMI (primary outcome), post-conception fasting glucose (secondary outcome), and neonatal adiposity, and will improve other maternal, prenatal, and neonatal health outcomes.

Material and methods

Overview.

This two-arm, parallel randomized clinical trial (clinicaltrials.gov identifier NCT05674799) will include 360 females with overweight/obesity and moderate-to-high likelihood of pregnancy within 24 months [23]. We will randomize them 1:1 to receive the 12-month, pre-conception lifestyle intervention or usual care. We will assess clinical and behavioral outcomes at enrollment, conception, throughout pregnancy, and delivery for those who become pregnant within 24 months of enrollment, and measure neonatal adiposity and other offspring outcomes at birth.

Setting and team structure.

This study will be conducted at the University of Colorado Anschutz Medical Campus and Denver Health Medical Center, the latter being a safety net hospital serving the Denver metro area. The principal investigators and co-investigators provide scientific oversight to the trial. Project managers at each site oversee day-to-day operations of the study, including all research assistant activities. Research assistants at both sites are responsible for participant recruitment, enrollment, and data collection. Participants will be recruited from Denver Health, and Denver Health staff will deliver the intervention and control programs. The Colorado Multiple Institutional Review Board approved the study protocol. A data and safety monitoring board will review safety and study progress at 6- to 12-month intervals.

Participants.

We will recruit 360 participants. Inclusion criteria are: established patient at Denver Health; biologically female (inclusive of all gender identities); aged 18–39 years; English- or Spanish-speaking; BMI ≥25 kg/m² (≥23 kg/m² if study participant self-identifies as Asian); self-report of engaging in activities that lead to pregnancy in the past 3 months and/or the possibility of pregnancy within 24 months (e.g., currently trying to become pregnant; currently using contraception but planning to discontinue within 24 months; not planning pregnancy and not using contraception but currently/expecting sexual activity with a biological male partner or other fertility options). Exclusion criteria are: currently pregnant; history of non-gestational diabetes (type 1, type 2); use of long-acting contraceptives (intrauterine device, implant, injection) with plans to continue for ≥2 years (barrier & short-acting hormonal contraception permitted given ease of discontinuing to facilitate pregnancy); previous sterilization procedures (e.g., tubal ligation, hysterectomy); documented infertility or unsuccessfully trying to conceive for ≥12 months; or prior participation in the NDPP.

We will identify potential participants from a patient registry based on medical records (e.g., patients meeting age, biological sex, BMI criteria, etc.), as well as provider- and self-referrals. We will conduct additional pre-screening via chart reviews for last known family planning status. Staff will contact potentially eligible individuals by phone to explain the study, confirm eligibility, and schedule an enrollment visit as indicated. Staff will obtain written informed consent from participants at enrollment. Participant characteristics are expected to match females <40 years in our prior NDPP cohorts (Table 1). Electronic data will be kept in secure, password-protected servers and computers, and paper data will be kept in locked filing cabinets in locked offices to protect participant confidentiality.

Table 1.

Anticipated characteristics of enrolled participants based on background population

Characteristic	% anticipated
Race and ethnicity
Latinx	73%
Non-Hispanic Black	14%
Non-Hispanic White	9%
Other races and ethnicities	4%
Income <133% of federal poverty level	75%
Mean baseline BMI (kg/m2)	36
Primary diabetes risk factor
HbA1c of 5.7–6.4%	69%
History of GDM	14%
Positive screen on diabetes risk test[28]	17%

Intervention condition.

The evidence-based NDPP promotes lifestyle change to lower risks for type 2 diabetes [24]. The NDPP curriculum is disseminated by the US Centers for Disease Control and Prevention (CDC), is available from registered delivery organizations in all states nationally [25], is cost-effective [26], and has expanding insurance coverage [27] to support broad uptake. The CDC provides delivery guidelines that allow flexibility to better serve target populations [28] and for continued participation upon pregnancy [24]. The program consists of 25 hour-long, group sessions (16 sessions in months 1–6 occurring every 1–2 weeks; 9 sessions in months 7–12 occurring every 2–4 weeks). For this trial, we follow these delivery guidelines and plan to launch eight cohorts of participants (5 cohorts in English, 3 in Spanish, per population need). As such, each cohort in the intervention arm will include ~22–23 participants to attend the yearlong NDPP together. Bilingual Lifestyle Coaches will receive CDC-approved training to lead the NDPP sessions. Lifestyle Coaches are lay healthcare professionals.

For this trial, we developed an enhanced version of the NDPP (NDPP-NextGen) with four primary enhancements for females likely to become pregnant: First, we will incorporate content on pre-conception/prenatal health into the NDPP curriculum and target 3% weight loss and normoglycemia before pregnancy, based on our successful preliminary data [29]. Participants will be encouraged to make sustainable improvements in diet and activity, but without pre-set goals to better accommodate a diverse, low-income population [30]. Upon pregnancy, guidance will support appropriate GWG, breastfeeding, and postpartum weight loss, based on current recommendations [31–33] and with continued obstetric care. Second, we will provide motivational “pre-sessions” that consist of an initial hour-long group session preceding the first official NDPP group session to promote increased knowledge of diabetes risks, self-efficacy, and readiness to change. Pre-sessions were successfully piloted at Denver Health [34] and are now recommended by the CDC [28]. Pre-session content will be tailored to further discuss how preconception weight loss can support healthy pregnancies and offspring. All NDPP-NextGen participants will be invited to attend the pre-session. Third, we will deliver the NDPP-NextGen sessions remotely via phone- and/or video-conference, which facilitates participation regardless of typical barriers to in-person attendance [35]. All NDPP-NextGen participants in a cohort will join the same remote meeting. Fourth, all NDPP-NextGen participants will be biological females <40 years of age (rather than biological males and females of all ages, as is usual practice) to increase personal relevance, promote social cohesion, and allow for added content on preconception/prenatal health [35]. These enhancements align with the Health Belief Model, in which behavior is determined by perceived seriousness, susceptibility, benefits, and barriers, as well as cues-to-action, self-efficacy, and modifying variables such as personal characteristics [36]. For example, pre-sessions may increase perceived seriousness, susceptibility, and benefits; remote delivery addresses perceived barriers; and contenttailored for the target population is responsive to personal characteristics.

Control condition.

A usual care control arm is necessary to compare outcomes associated with NDPP-NextGen to potential changes without a targeted, intensive intervention (e.g., slight increase or decrease in BMI [37, 38]). We did not include a contemporary control arm of standard NDPP given we have previously shown that the standard NDPP insufficiently engages young females [39]. Rather, participants randomized to the control arm will receive usual care for females of childbearing potential with overweight/obesity, including annual preventive visits to discuss family planning or health priorities (including weight-related issues), and “sick” or other encounters as needed. We will also provide educational handouts that Denver Health routinely distributes about healthy lifestyle change before, during, and after pregnancy, with instructions to follow up with their providers as needed. We expect the usual care arm to receive minimal, if any, clinical weight loss services during the study period [40]. Usual care participants who receive weight loss services will be retained in the analysis, per intention-to-treat; however, staff will ask participants about receipt of such services (including NDPP services that may be available in the community, bariatric surgery, and weight management pharmacotherapy) at all research visits to assist with interpretation of results.

Data collection.

In-person research visits for data collection will occur at enrollment, post-conception (<8 weeks gestation), and at delivery (before discharge). Remote research visits will occur in late pregnancy (28–32 weeks gestation). The schedule and outcomes of data collection are summarized in Table 2. Participants will be compensated up to $300 for completing all research visits, and receive transportation support (taxi, parking fees) as needed. Unless otherwise specified, data will be collected via surveys administered in Research Data Capture (REDCap).

Table 2.

Summary of clinical and behavioral outcomes

Outcomes	Methods of assessment	Enrollment	Conception	<8 weeks gestation	28–32 weeks gestation	Delivery
BMI post-conception (primary outcome)	In-person measures of weight, height	•		•
BMI around conception	Home cellular-enabled scale weight (weekly)	•	•	•	•	•
Glycemia post-conception (secondary outcome)	Fasting glucose, A1c	•		•
GDM in mid-pregnancy	Medical records				•
GWG (total, per IOM guidelines, timing/rate)	Home scales, medical records		•	•	•	•
Diet (energy, macronutrients, micronutrients, quality)	Automated Self-Administered 24-hour recall	•		•	•
Physical activity (all types and intensities)	Pregnancy Physical Activity Questionnaire	•		•	•
Smoking and tobacco use	Self-report recent and lifetime use	•		•	•
Self-efficacy for weight management	Weight Efficacy Lifestyle Questionnaire- Short Form	•		•	•
Depressive symptoms (for safety monitoring)	Edinburgh Postnatal Depression Scale	•		•	•
Other maternal outcomes (preeclampsia, etc.)	Medical records			•	•	•
Neonatal adiposity	PEA POD air displacement plethysmography					•
Other neonatal outcomes (birthweight, LGA, etc.)	Medical records					•

Demographics (age, race, ethnicity, language, education, income, insurance, household composition), reproductive history (gravidity, parity) and diabetes risks (prediabetes laboratory results, prior GDM, diabetes risk survey score [28]) will be obtained from medical records during pre-screening and verified with participants at enrollment.

For conception monitoring, staff will review medical records and outreach monthly to ask about menstrual cycle and possible pregnancy. Staff will also provide home pregnancy tests to facilitate early detection of conception. Participants are also encouraged at enrollment and at each check-in to contact staff as soon as they become aware of conceiving. When participants report conception, an in-person visit will be scheduled promptly, targeting <8 weeks gestation, to confirm pregnancy and collect post-conception data.

Maternal post-conception BMI (primary outcome) will be calculated using total weight assessed via a high-capacity, medical-grade scale. Staff will take two weights (and a third if differing by >0.5kg; averaging the closest two for analysis), while participants are in light indoor clothing without shoes. Participant height at enrollment and post-conception will be measured by staff with a wall-mounted stadiometer after participants remove shoes. Two measurements will be taken at each time, with a third taken when measures differ by >0.5 cm, and the closest two will be averaged for analysis. We will also assess BMI around the time of conception with weights obtained from cellular-enabled scales (BodyTrace^®). Participants will use these scales at home to assess morning fasted weight in light clothing at least weekly, prompted by staff as needed. Scales automatically transfer weights to a HIPAA-compliant database that is accessible only by the research team. We will average the two weights closest to the estimated date of conception for analysis. Home weighing will continue to measure GWG throughout pregnancy. We will also abstract all weights from prenatal medical records as an alternative method to assess GWG.

Maternal post-conception glycemia will be evaluated as 8-hour fasting glucose (secondary outcome) collected with venipuncture sampling at the post-conception study visit. We will simultaneously measure hemoglobin A1c given its clinical relevance and higher utility than early pregnancy fasting glucose [41]. We will abstract GDM screening results and treatment from medical records, as Denver Health follows the recommended 2-step screening and diagnostic procedure at 24–28 weeks [42].

Maternal behavioral outcomes will be assessed at enrollment, post-conception, and mid-pregnancy. Diet will be assessed with the Automated Self-Administered 24-hour Dietary Recall system [43]. Two recalls will be conducted at each timepoint, with the first completed in-person with staff assistance and the second completed remotely within ~10 days. We will use a measurement error model to estimate usual dietary intake at each timepoint [44], and derive daily profiles of macronutrient, micronutrients, food groups, and diet quality [45]. We will assess physical activity with the Pregnancy Physical Activity Questionnaire [46], while adjusting metabolic task equivalents for pregnancy [47]. Items are relevant for females prior to pregnancy and thus facilitate consistent data collection. We will calculate average activity intensity (sedentary to vigorous) and type (household, occupational, exercise, transportation) over the past 3 months. Additional surveys will evaluate tobacco and marijuana use, weight management self-efficacy [48], body size perception [49], and depressive symptoms [50]. Further preconception and prenatal data will be abstracted from medical records: date of conception (early ultrasound dating, self-reported first day of last menstrual period), miscarriage, stillbirth, fetal death, pre-term birth, hypertensive disorders of pregnancy, mode of delivery, and use of clinical services before and during pregnancy.

The primary offspring outcome is neonatal adiposity (percent of total mass that is fat) assessed via air displacement plethysmography within 48 hours of birth. We will use the PEA POD (Cosmed, Italy), which is specially designed to measure infant body composition [51]. Body mass and volume measurements are used to calculate body density, and fat and fat-free mass density coefficients [52] are used to obtain estimates of fat-free mass (g), fat mass (g), and adiposity (percent of total mass that is fat). We will take two measurements, and a third if adiposity differs by >2% (averaging the closest two). Other neonatal data will be abstracted from records on gestational age, birthweight, macrosomia (>4000g), low birthweight (<2500g), size per gestational age (large >90^th percentile, small <10^th percentile), birth trauma (fractures, brachial plexus injuries, skull/head bleeds), shoulder dystocia, intensive care unit admission, major congenital anomaly, and hypoglycemia.

Biospecimens (maternal blood, urine, and stool at each in-person visit, and offspring cord blood at delivery) will be collected and stored for future ancillary studies related to maternal-child health. Participants will opt in or out of biospecimen banking during the consenting process at enrollment.

We will also collect process evaluation data to assess NDPP-NextGen implementation. Reach and representativeness of participants relative to the target population of reproductive-aged females with overweight and obesity will be evaluated via recruitment data, including the number and percent of potentially eligible individuals who are identified and enroll, as well as reasons for non-enrollment. Dose will be assessed by participant-level attendance tracking, including dose prior to and during pregnancy. We will also assess mode of attendance at each session (joining by video-conference or phone-only). Fidelity will be examined by observing 10% of sessions for adherence to core components, while tracking overall observations (e.g., group interactions), to be used for quality control and staff re-training as needed. Observations will be considered alongside quantitative data to inform results interpretation.

Randomization.

We will randomize eligible, consenting participants to NDPP-NextGen or usual care in a 1:1 ratio. We will use a permuted block randomization approach with randomly chosen block sizes of 2, 4, or 6. The approach balances randomization within time, but prevents study staff from guessing randomization assignment. The randomization scheme will be developed by the study statistician, who will remain blinded to assignment until completion of analysis. Assignment will not occur until after baseline measures are completed to maintain participant and staff blinding during baseline data collection. Data will remain blinded to study investigators until data collection is complete. Quality control and progress reports will be presented to study staff under the blind.

Statistical analyses.

Data will be analyzed in SAS (v9.4, Cary, NC) by modified intention-to-treat, as only those participants who conceive can be included in analyses of weight, glycemia, and behavioral outcomes from conception onward. Participants with missing data for any outcome will be excluded for that particular analysis, as last-value-carried forward calculations may be biased [53]. Participants who experience pregnancy loss will be retained in analyses for which data are available. For all analyses, we will compare the demographics of excluded participants to those retained to assess differential dropout. No interim analyses or stopping rules are planned. Statistical significance is set as two-sided α<0.05 for all outcomes.

We will use linear and logistic regression for continuous and categorical outcomes, respectively, constructing separate models for each outcome. We will use a Wald test with Kenward-Roger [54] degrees of freedom to evaluate the effect of a treatment indicator variable for NDPP-NextGen versus usual care as the predictor. We will construct unadjusted models, and then models adjusted for maternal age, race, ethnicity, language, baseline value (e.g., BMI at enrollment for BMI analyses), and cohort (1–8). We will not test for interactions of race, ethnicity, or language with treatment because we have previously reported no differences in outcomes based on these characteristics [55]. For offspring outcomes, we will additionally adjust for gestational age at birth and infant sex. Because neonatal adiposity differs by sex [56], we will consider the interaction between infant sex and treatment arm; if not statistically significant, it will be removed from the model and only main effects will be interpreted. To model the variance, we will use a random effect for cohort in the intervention group. We will use an unstructured variance model for repeated measures within person across time. Accounting for both repeated measures across time and clustering within the intervention group produces a Kronecker product compound symmetric covariance structure [57]. We will verify assumptions to confirm model validity, and report parameter estimates, errors, test statistics, degrees of freedom, 95% confidence intervals, and p-values.

Mediation analyses will determine which marker of maternal health explains the greatest amount of variation in neonatal adiposity after intervention, which can inform strategies to maximize the impact of preconception intervention. For example, if peri-conceptional glycemia is a stronger mediator of offspring adiposity than maternal GWG, future strategies should focus more on controlling maternal glucose levels rather than GWG [30]. We will explore mediators for all maternal peri-conceptional and prenatal outcomes that vary by treatment assignment [58], beginning with neonatal adiposity. We will confirm the association of the mediating variable with the outcome variable using linear or logistic regression, and then examine the attenuation of this association after adjusting for the mediator. We will interpret the statistical significance of observed mediation with the bootstrap method [59], and consider mediators to be important if they attenuate treatment effects by ≥25%.

We will conduct a sensitivity analysis to estimate possible bias induced by analyzing only participants who conceive. After randomization, we will match NDPP-NextGen and usual care participants 1:1 using a propensity score based on age, BMI, and gravidity (variables known to affect conception) [60]. This ultimately creates four pair types: both participants conceive, neither conceive, only treatment conceives, only control conceives. We will calculate BMI at conception within pair types based on home scale weights. When only one participant conceives, we will use her time from baseline to conception to define weight from baseline for the non-conceiving participant. When neither conceive, we will randomly select a time to conception from another pair to assign a pseudo-conception time from baseline. Matching creates correlation within each matched pair. We will fit a mixed model with the two paired BMI values as the outcome, randomization arm and propensity score as the fixed effect predictors, and a random effect for each pair. We will evaluate the difference in BMI between the two randomized groups, and compare it to the result from our primary analysis. If they are similar (i.e., confidence intervals overlap), we can conclude trial results were not substantially influenced by selection bias.

Power calculations.

We will enroll 360 participants, anticipating that ≥60% (n=172) will conceive within 24 months, as previously demonstrated. Our preliminary data provided effect size estimates for expected change in BMI and A1c, which we use as a proxy for early pregnancy hyperglycemia because data on fasting glucose in early pregnancy was unavailable [29]. We expect BMI at conception to be ≥1.7 kg/m² lower and early pregnancy A1c to be ≥0.32% lower in NDPP-NextGen than usual care. More conservatively, if BMI and A1c remain stable in control participants (vs increasing as in our preliminary data), we expect effect sizes of −0.7 kg/m² and −0.3%, respectively.

Power was computed in GLIMMPSE (version 3) [61] assuming a linear mixed model with clustering by cohort. We used data from prior NDPP participants at Denver Health as input estimates for variances in outcomes (n=1170 and n=27 females aged <40 years participating in the NDPP with BMI and early pregnancy A1c data, respectively) and correlations within cohorts. We assumed 80 conceiving participants per treatment arm. We expect correlated outcomes between participants in the same NDPP-NextGen cohort but not in the control arm, yet conservatively assumed clustering in both study arms. With the Wald test and 0.05 Type 1 error rate, we estimate >0.90 power to detect differences in BMI and A1c among 160 conceiving participants (Figure 1). We also expect >0.90 power to detect significant treatment effects in fasting glucose (our a priori secondary outcome), as fasting glucose is more sensitive than A1c for capturing early dysglycemia [42, 62].

Figure 1.

Power for post-conception maternal BMI (left) and glycemia (right).

Data and safety monitoring plan.

Individual participant survey responses and biological results will be reviewed on a rolling basis for quality control and safety. Glucose and A1c results will be returned to participants as soon as they are available. When safety thresholds are exceeded for glucose (≥126 mg/dl), A1c (≥6.5%), or weight (loss or plateau during pregnancy), results will be reviewed by the study OB/GYN (Fabbri) and participants referred for follow-up clinical care. Research staff will monitor depression survey responses at the study visit (and at least weekly for asynchronously collected data), and contact emergency services (i.e., 988) if a participant discloses imminent risk of harm to self or others.

A Data and Safety Monitoring Board consisting of 4 external individuals with expertise in obesity and diabetes around pregnancy, neonatology, and biostatistics was appointed to oversee the study. The Board will meet every 6–12 months to review safety and study progress (enrollment progress, number of research visits conducted, attrition and accompanying reasons, completeness of data, proportion of women conceiving, proportion of women with A1c ≥6.5% at research visits, adverse events, and severe adverse events). The Board may request to review unblinded results in a session closed to the investigators to determine if safety is related to group assignment. The Board will determine whether adverse event rates are consistent with pre-study assumptions and whether the study is on track to be completed and accomplish the stated aims.

Results

Recruitment is underway as of December 2022. Initial participants assigned to NDPP-NextGen will begin the intervention in summer 2023. Data analysis and results reporting is expected to be completed in 2027.

Discussion

Reducing maternal obesity/diabetes risks at conception and early pregnancy, especially in diverse and underserved females, is critical to breaking the cycle of disease across generations. Research shows that lifestyle interventions may reduce obesity and diabetes risks; however, no studies have rigorously evaluated effects of a scalable lifestyle intervention for diverse and under-resourced females beginning before conception. A randomized clinical trial is needed to evaluate effects of a scalable lifestyle intervention on maternal-child health in a high-risk population.

Our preliminary data suggest the widely available NDPP can safely lower maternal weight and glycemia in a diverse, underserved sample [29]. We previously identified promising strategies to boost engagement of young females in the NDPP [35]. We tested how to pre-screen via medical records and then verify current family planning status to achieve ≥60% conception by 24 months. We included other innovative elements in this trial: targeting modest weight loss (3%) to improve glycemia [30]; cellular-enabled home scales to assess weight at conception; and measuring neonatal adiposity in addition to birthweight. This study is powered to detect clinically relevant effects, as every 1 kg/m² increase in pre-pregnancy BMI is associated with a 14% increase in GDM risk [63], 8% increase in gestational hypertension [64], and a 33% reduction in LGA births [65]. Further, the anticipated effect on A1c corresponds to a 9 mg/dl shift in average glucose levels [66], which is associated with a 10% increased risk of LGA birth [67] and a 0.03-unit increase in childhood fat mass index [68].

We acknowledge likely limitations of this trial. Participants may be challenged to accurately self-report likelihood of pregnancy, which could result in fewer pregnancies within 24 months of enrollment. Yet if we observe the same effect size as in our preliminary data (1.7 kg/m² difference in BMI and 0.32% difference in A1c) rather than the conservative effects projected here (0.7 kg/m² difference in BMI and 0.30% difference in A1c), we could detect statistically significant effects in a smaller sample. We expect lower power for additional outcomes (e.g., gestational diabetes, macrosomia) and the proposed mediation analyses. Recruitment is limited to a single health care system. Our study requires initial contact by phone to enroll and phone/internet service to participate; thus, we may be systematically missing individuals who do not have a phone. Results will reflect females enrolled in a conception-related research study, who may differ from females unwilling to participate. Interpretation of A1c results may be limited by the normal downward shift in levels beginning at 10–20 weeks gestation of A1c [69] and genetic variations in hemoglobin relevant to our diverse sample [70].

Conclusions

This clinical trial has high potential impact because the intervention could be disseminated among the >1800 organizations currently delivering the NDPP [25]. Future research can include following offspring to assess continued effects on childhood health outcomes. Implementation science research could assess the possibility of implementing NDPP-NextGen broadly, with the goal of extending potential health benefits widely for generations to come.

Abbreviations

BMI

body mass index

CDC

Centers for Disease Control and Prevention

GDM

gestational diabetes

GWG

gestational weight gain

NDPP

National Diabetes Prevention Program

REDCap

Research Data Capture

US

United States