Intervening during and after pregnancy to prevent weight retention among African American women.

Efforts to prevent postpartum weight retention in extant clinical trials of African American women have proven exceedingly challenging. The primary purpose of this pilot study was to determine whether a behavioral intervention implemented in early pregnancy through 6 months postpartum could increase the proportion of African American women who were at or below their early pregnancy weights by 6 months postpartum. We additionally evaluated whether mothers' postpartum weight loss could be maintained at 12 months postpartum. Participants were 66 socioeconomically disadvantaged African American women (36% overweight, 64% obese) randomly assigned to a behavioral intervention or usual care group. The intervention, implemented from early pregnancy to 6 months postpartum, promoted weight control through: (1) empirically supported behavior change goals; (2) interactive self-monitoring text messages; (3) weekly to monthly health coach calls; and (4) skills training and support through Facebook. In modified intent-to-treat analyses, participants assigned to the intervention were significantly more likely to be at or below their early pregnancy weights by 6 months postpartum compared to usual care (56% vs. 29%, p = 0.04). At 12 months postpartum, the maternal weight difference between intervention and usual care groups was not maintained (41% vs. 38% respectively at or below early pregnancy weights, p = 0.83). Findings suggest that a combined pregnancy and postpartum weight control intervention improves 6 month weight outcomes in socioeconomically disadvantaged African American women with obesity. Longer interventions may be needed to overcome late postpartum weight gain among this high risk group. Clinical trial registration number: ClinicalTrials.gov identifier NCT01530776.

RCT Entities:   Population Interventions Outcomes

Introduction

Returning to early pregnancy weight by 6 months postpartum is particularly challenging for African American women with obesity (Headen et al., 2012). Data suggest that African American women retain twice as much weight as than their white counterparts between 6 and 18 months postpartum, even after adjustment of important covariates such as pre-pregnancy body mass index (BMI) and parity (Boardley et al., 1995, Keppel and Taffel, 1993, Parker and Abrams, 1993).

The strongest predictor of postpartum weight retention is the amount of weight gained in pregnancy (Gunderson and Abrams, 1999), yet only one publication to our knowledge has evaluated the combined effect of a pregnancy and postpartum weight control treatment on 6 month weight outcomes. The findings from this study were promising (Huang et al., 2011); compared to participants randomized to postpartum treatment alone or to a control group, participants randomized to the pregnancy and postpartum intervention retained significantly less weight at 6 months postpartum, with significantly fewer participants retaining > 5 kg at follow-up (18% in the combined treatment group vs. 42% in postpartum only treatment vs. 52% in the control group, p < 0.001). However, this study was conducted in a small sample of Taiwanese mothers whose average BMI at baseline was 21 kg/m2, limiting generalizability to populations at highest risk for postpartum adiposity.

The primary purpose of this pilot study was to determine whether a behavioral intervention implemented in early pregnancy through 6 months postpartum could increase the proportion of obese African American women who were at or below their early pregnancy weights by 6 months postpartum. We additionally evaluated whether mothers' postpartum weight loss could be maintained at 12 months postpartum.

Methods

Study participants were recruited by trained research staff at 2 large outpatient obstetric offices at Temple University between 2013 and 2014. Using Temple's electronic medical record (EMR), study staff identified potential participants by BMI, gestation length, and age, and then approached them in waiting rooms to assess trial interest. Inclusion criteria were: 1) age ≥ 18 years; 2) self-identification as African American; 3) gestation length < 20 weeks; 4) first trimester BMI 25–45 kg/m2; 5) Medicaid recipient (income proxy); 6) cell phone ownership with unlimited text messaging; and 7) Facebook member. Women that endorsed current tobacco use, were carrying multiples, or reported conditions that required specialized nutritional care were excluded. A baseline assessment was scheduled and completed at our research center (in a building separate from our obstetric practices) after which participants were randomly assigned by computer-generated numbers into the intervention or usual obstetric care. Additional assessments at 36 weeks' gestation, 6 months postpartum and 12 months postpartum were performed; participants who attended all visits received a total of $120 for time/travel. Pregnancy findings were published previously (Herring et al., 2016). All subjects gave written informed consent for participation, which was approved by the Institutional Review Board at Temple University.

Participants randomized to the intervention arm received a behavioral lifestyle intervention designed to: 1) prevent excessive weight gain in pregnancy (delivered from baseline to 36 weeks' gestation); and 2) promote weight loss postpartum (delivered between 10 weeks' and 6 months after childbirth). During our formative work, we identified and prioritized a series of behavior change goals that were evidence based, relevant to the patient population, and could be easily self-monitored through text messaging. Examples included (as provided to participants): “Limit junk and high fat foods to no more than 1 per day”; “Limit sugary drinks like juice and soda to no more than 1 per day”; “Walk 5000 steps daily”; “Weigh yourself weekly”. While goals were nearly identical in pregnancy and the early postpartum period, framing of messages differed (e.g., pregnancy messages were framed around “being healthy for baby”, while postpartum messages were focused on “being a role model”). Skills training and support were delivered via three duplicate mechanisms during and after pregnancy: 1) daily skill-building text messages tailored to each behavioral goal; 2) weekly Facebook posts with links to websites and videos; and 3) weekly to monthly 15-minute scripted calls with a bachelor's level health coach so to problem solve barriers, provide support, and build participant self-efficacy for behavior change. We provided information about cost and community resources to heighten salience of materials. Self-monitoring texts were additionally sent to participants 3–4 times weekly to probe about behavioral adherence. Text message prompts in the morning (e.g., “Please text us # cups of sugary drinks u had yesterday”) were followed by immediate automatic feedback to reinforce successes and/or offer support (e.g., “U had 2 sugary drinks. Ur almost at ur goal. Here's a tip: add sugar-free flavoring to ur water to give it more taste”). To aid in goal attainment, intervention participants received a digital scale for self-weighing, pedometer to track steps, water bottle, and portion plate to encourage smaller portions. A binder with print versions of program content was also provided for skills training if technology access was lost.

The primary outcome for this study was the proportion of women at (within 0.9 kg) or below their early pregnancy weights by 6 months postpartum (Phelan et al., 2011). Secondary outcomes included proportion of women at or below their early pregnancy weights by 12 months postpartum, intervention engagement (e.g., the proportion of participants who responded to ≥ 50% of the self-monitoring text prompts and number of coach calls completed), and treatment acceptability assessed via survey at 6-month postpartum follow-up. Weights were assessed at Temple's obstetric practices in pregnancy and then abstracted from the EMR; we used earliest abstracted prenatal weight to define early pregnancy weight and additionally abstracted last measured weight prior to delivery to calculate total gestational weight gain. We and others have found that the absolute agreement between research and prenatal clinic measured weights is extremely high, providing strong support for exchangeability (Vesco et al., 2014, Leo et al., 2014). Six-month postpartum weights (end of treatment) and 12-month postpartum weights (6-month follow-up) were assessed via calibrated scales at our research center during in-person visits. Missing postpartum weights at 6 months (n = 3) and 12-months (n = 4) were additionally filled with EMR data (weights were accepted ± 30 days of planned 6-month and ± 60 days of planned 12-month postpartum assessments).

To compare baseline characteristics and outcome data between the intervention and usual care groups, we used Pearson chi-square or Fisher's exact tests for categorical variables, and t-tests or Mann-Whitney U tests for continuous variables. A multivariable logistic regression model was used to determine the effect of treatment group on our primary outcome, adjusting for potential confounders of postpartum weight retention, including early pregnancy BMI, parity, and maternal age. Our a priori analysis plan followed a modified intent-to-treat approach (ITT) that conservatively assumed participants who were lost to follow-up were treatment failures (e.g., were not at or below their early pregnancy weights), an approach that has been published elsewhere (Phelan et al., 2011). We also explored whether gestational weight gain (last measured weight before delivery minus first measured weight in early pregnancy) was predictive of 6-month and 12-month postpartum weight change using multiple linear regression.

Results

Sample

Fig. 1 summarizes the participant flow. The intervention and usual care groups did not differ significantly with respect to baseline characteristics (Table 1).

Fig. 1

CONSORT flow diagram.

Table 1

Maternal demographic characteristics and weight at baseline, delivery, 6-months postpartum, and 12-months postpartum.

	Usual care(n = 33)mean (SD) or %	Intervention(n = 33)mean (SD) or %	p value
Baseline
Maternal age (years)	25.0 ± 5.7	25.9 ± 4.9	0.34
Early pregnancy weight (kg)	87.2 ± 14.0	86.7 ± 15.6	0.90
Early pregnancy body mass index (kg/m2)	32.2 ± 5.4	33.5 ± 5.8	0.50
Nulliparous	30%	27%	0.79
Single	67%	73%	0.59
Unemployed	52%	55%	0.81
Education			0.63
Some high school or less	15%	21%
High school graduate	61%	49%
Technical school	12%	9%
Some college or more	12%	21%
Timing of weight measures (weeks' gestation)	7.7 ± 4.0	7.9 ± 3.4	0.79
Deliverya
Exceeded IOM gestational weight gain guidelines	66%	37%	0.03
Last measured weight before delivery (kg)	101.7 ± 12.6	95.7 ± 15.5	0.11
Timing of weight measures (weeks before delivery)	0.6 ± 0.8	0.5 ± 0.4	0.60
6-months postpartum
Participants at or below early pregnancy weightb	29%	56%	0.04
6-month postpartum weight (kg)c	92.0 ± 13.5	87.5 ± 12.6	0.20
Timing of weight measures (months' postpartum)	6.2 ± 0.7	6.1 ± 0.5	0.42
12-months postpartum
Participants at or below early pregnancy weightd	38%	41%	0.83
12-month postpartum weight (kg)c	92.3 ± 14.6	89.6 ± 14.8	0.52
Timing of weight measures (months' postpartum)	12.3 ± 0.7	12.5 ± 0.8	0.41

Delivery analysis excluded participants with pregnancy loss or preterm birth (n = 4 usual care, n = 6 intervention). IOM = Institute of Medicine.

On the basis of modified intent-to-treat analyses, conservatively assuming the 2 participants lost to follow-up did not return to baseline. We additionally excluded participants with pregnancy loss or those that became pregnant (n = 2 usual care, n = 6 intervention).

On the basis of completer analyses.

On the basis of modified intent-to-treat analyses, conservatively assuming the 6 participants lost to follow-up did not return to baseline. We additionally excluded participants with pregnancy loss or those that became pregnant (n = 4 usual care, n = 6 intervention).

Maternal weight outcomes

Consistent with what we reported previously (Herring et al., 2016), participants assigned to the intervention group were significantly less likely to exceed weight gain guidelines in pregnancy compared to usual care (37% vs. 66%, p = 0.03; Table 1). By 6 months postpartum, modified ITT analyses revealed that intervention participants were significantly more likely to be at or below their early pregnancy weights compared to those in usual care (56% vs. 29%, p = 0.04; Table 1). Similar results were observed in adjusted analyses (odds ratio [OR]: 3.5; 95% confidence interval [CI]: 1.0, 11.8, p = 0.04). Of note, treatment effect was reduced when gestational weight gain was entered into the model, suggesting potential mediation (OR: 2.8, 95% CI: 0.8, 9.8, p = 0.11).

By 12 months postpartum, no differences were found in the proportion of participants at or below their early pregnancy weights between intervention and usual care groups (41% vs. 38% respectively, p = 0.83). While we found that gestational weight gain was positively associated with postpartum weight change at 6 months, independent of potential confounders and group assignment (b = 0.6 kg, 95% CI: 0.2, 1.0, p = 0.006), we did not find the same to hold true at 12 months postpartum (b = 0.3 kg, 95% CI: − 0.3, 0.9, p = 0.30).

Intervention engagement and treatment acceptability

The majority of participants in pregnancy (70%) and postpartum (68%) periods responded to ≥ 50% of the self-monitoring text prompts. Intervention participants also completed an average of 4 ± 1.5 coaching calls (expected calls = 7) during the first 12 program weeks in pregnancy as previously reported (Herring et al., 2016), and an average of 3 ± 1.7 coaching calls (expected calls = 6) postpartum. Among intervention participants who completed the postpartum treatment acceptability questionnaire (n = 22; 79%), 87% reported the skills they learned in the program were extremely useful (at least an 8 on a 10-point scale); 91% found the text messages and 87% found the coach calls extremely useful.

Discussion

In this study, we found that a combined pregnancy and postpartum weight control intervention improved 6 month weight outcomes in socioeconomically disadvantaged African American women with obesity. Given the dearth of efficacious postpartum weight loss studies in ethnic minority women, our findings suggest that including pregnancy in the intervention design may be an important strategy to prevent postpartum weight retention among this high risk population. Once treatment ended, however, the maternal weight differences between intervention and usual care groups were not maintained at 12 months postpartum.

The 6-month weight findings of our study are similar to Huang's findings in Taiwanese women (Huang et al., 2011), who also observed reductions in weight retention at 6-months postpartum following weight control treatment in pregnancy and the early postpartum periods. Additionally, both trials had significant reductions in pregnancy weight gains that likely mediated intervention effects (at least in part) at 6-months postpartum (Herring et al., 2016, Huang et al., 2011). However, by 12-months postpartum, gestational weight gain no longer predicted weight retention in our study, and thus, African American women randomized to the intervention group were unable to maintain their weight loss. Longer interventions may be needed to overcome late postpartum weight gain among this high risk group.

Intervening in pregnancy alone, however, may be unlikely to produce the same magnitude of weight loss we found by 6 months postpartum. In one of the largest weight control interventions implemented solely in pregnancy, Phelan et al. (2011) reported that 25.6% of participants with overweight or obesity returned to their preconception weights (within 0.9 kg) or below at 6 months postpartum. We observed more than double this proportion (56%) in our study. This finding highlights the need to intervene both during and after pregnancy if the goal is to substantially reduce postpartum weight retention.

Strengths of our study include its randomized design, treatment in pregnancy and the postpartum periods, focus on African American women with obesity (an understudied group at high risk for weight retention), use of mobile phone technologies for intervention delivery, and duration of participant follow-up. However, any conclusions must be tempered by study limitations including the pilot nature and small sample size. Our design did not allow for isolation of discrete intervention components, and thus, we were unable to determine which component attributed to the observed effects. We used clinic measured weights in calculations of some outcomes, and thus, cannot be certain these weights were collected with the same degree of quality (e.g., calibrated scales, no shoes) as weights collected by research staff. Clinic staff were blind to randomization assignment; however, so any measurement bias would likely be non-differential across treatment groups. Finally, while we recognize that the modified ITT approach has its limitations, analyses that used data of completers yielded near-identical results.

In conclusion, the implementation of a behavioral intervention during and after pregnancy was effective in increasing the proportion of obese African American women who were at or below their early pregnancy weights by 6 months postpartum. These promising findings should trigger a larger definitive study in this and potentially other populations.

Funding

This study was supported by grants from the National Institutes of Health (NIH K23 HL106231) and the Health Resources and Services Administration (HRSA R40MC26818) of the U.S. Department of Health and Human Services (HHS). The information or content and conclusions of this study are those of the authors and should not be construed as the official position or policy of, nor should any endorsements be inferred by NIH, HRSA, HHS or the U.S. Government.

Disclosures

At the time of the study, Dr. Herring served on scientific advisory boards for Novo Nordisk and Johnson and Johnson; Dr. Bennett served on the scientific advisory boards for Nutrisystem and the board of Scale Down; and Dr. Foster served on scientific advisory boards of Con Agra Foods, Tate and Lyle, and United Health Group. Currently, Dr. Foster is a full-time employee of Weight Watchers International. None of these entities have provided financial support for this study nor did they have any influence on the weight control methods in this study. All other authors declare no conflicts of interest.

Transparency document

Transparency document.