ABSTRACT
Background
Whether breastfeeding intensity, the relative amount of breastfeeding versus formula across the first year, relates to growth trajectories among infants exposed to gestational diabetes (GDM) in utero is unknown.
Objective
To identify growth trajectory patterns and associations with breastfeeding intensity among infants exposed to GDM.
Methods
SWIFT Offspring cohort (2009–2011) singleton, GDM‐exposed infants (n = 435) born at a Kaiser Permanente Northern California hospital. Monthly summed lactation intensity ratio (sumLIR, proportion of breastfeeding versus overall feeds over 7 days). Latent class mixed models for weight‐for‐age (WAZ), length‐for‐age (LAZ), and weight‐for‐length (WLZ) z‐scores, and skinfold thicknesses (triceps, subscapular, and sum of skinfolds). Multinomial logistic regression related sumLIR to growth trajectories.
Results
A higher sumLIR throughout the first year was associated with lower odds of class membership in trajectories with accelerated growth patterns for WLZ from early infancy (1‐mo OR: 0.11, 95% CI 0.02, 0.54) to 1 year (12‐mo OR: 0.84, 95% CI 0.75, 0.95), as well as a lower likelihood of having accelerated WAZ growth after 6‐months (12‐mo OR: 0.91, 95% CI 0.84, 0.98), but not skinfold trajectory classes.
Conclusions
Higher sumLIR was associated with more favourable growth trajectories, suggesting that continued breastfeeding, even when combined with infant formula, may benefit growth among GDM‐exposed infants.
Keywords: breastfeeding, gestational diabetes, lactation intensity, LCMM
Abbreviations
- FJ
100% fruit juice
- GDM
gestational diabetes
- KPNC
Kaiser Permanente Northern California
- LAZ
length‐for‐age z‐score
- LCMM
latent class mixed models
- LIR
lactation intensity ratio
- OGTT
oral glucose tolerance test
- sumLIR
cumulative monthly summed lactation intensity ratio
- WAZ
weight‐for‐age z‐score
- WLZ
weight‐for‐length z‐score
1. Introduction
Offspring born to a parent with gestational diabetes (GDM) are more likely to be larger [1] and have greater adiposity at birth [2], have unique growth patterns during infancy [3, 4, 5] and have a higher risk of overweight or obesity in early childhood [6, 7, 8]. Infants exposed to GDM are at higher risk for increased skinfold thickness and adiposity [9, 10] in mid‐childhood, which may be moderated by breastfeeding [11]. Yet, infants exposed in utero to GDM are less likely to be exclusively breastfed and more likely to be offered formula supplementation earlier [12].
In the general population, breastfeeding is associated with slower, more optimal growth in the first year of life and modestly reduced risk of obesity later in life [13, 14, 15]. Interestingly, breastfeeding is more strongly associated with differences in early growth and subsequent obesity in GDM exposed offspring [6, 16]. Among GDM exposed infants, exclusive breastfeeding for 6 months is associated with a lower weight‐for‐age z‐score (WAZ) at 14 months, lower odds of obesity at 1–5 years, and lower BMI‐for‐age at 2–5 years [17, 18]. Most research dichotomizes or trisects (full breastfeeding, mixed or full formula) breastfeeding and also assesses breastfeeding retrospectively, which reduces precision as there is variability on the breastfeeding‐formula continuum across infancy [19]. Prospective assessments with detailed collection of breastfeeding data offer an opportunity to examine breastfeeding intensity. Previously, in the SWIFT cohort, we found that more intensive breastfeeding, trichotomizing lactation intensity ratio in the first year, characterized as consistent, exclusive, or mostly breastfed, was associated with slower overall increases in WAZ and weight‐for‐length (WLZ) z‐scores independent of maternal obesity and GDM severity [20]. Our prior research leveraged linear regression methods that allow for assessing overall population mean growth for each breastfeeding category rather than time specific differences in velocity and curves with latent class mixed models (LCMM) that clusters among groups of individuals. LCMM is a data driven approach that identifies underlying growth trajectory patterns in the data and furthermore allows for evaluation of the associations among exposures that may coincide with differential effects on growth patterns [20]. While we have previously reported differences in growth and adipose tissue trajectory patterns between offspring exposed to GDM versus unexposed counterparts [1], the impact of breastfeeding and complementary feeding practices were not assessed. Consequently, we were not able to identify how potentially modifiable lifestyle factors such as breastfeeding and sugary beverage consumption during infancy differentially affect growth patterns for GDM exposed offspring [21].
This study sought to evaluate the role of monthly breastfeeding intensity and cumulative breastfeeding intensity on growth and subcutaneous adipose tissue patterns across the first year of life among GDM‐exposed infants. We hypothesized that a higher lactation intensity and no sugar‐sweetened beverages (SSB) or 100% fruit juice (FJ) exposures would be associated with more favourable growth trajectories accounting for maternal obesity and covariates.
2. Methods
2.1. Participants
The Study of Women, Infant Feeding and Type 2 diabetes after GDM pregnancy (SWIFT) included 1033 women with GDM to evaluate the association of breastfeeding intensity and duration with infant growth and adiposity in the first year of life, and has previously been described in detail [22, 23]. A subset of SWIFT mother‐infant dyads also enrolled in the SWIFT Offspring Study from 2009 to 2011 to measure infant anthropometry and adiposity. Briefly, women were eligible for the parent SWIFT study if they were diagnosed with GDM by the Carpenter and Coustan criteria [24], delivered a singleton birth ≥ 35 weeks gestation at a Kaiser Permanente Northern California (KPNC) hospital from 2008 to 2011 and met other study selection criteria [22]. For this analysis, participants were excluded if they were missing two or more measurements either at birth or the three research visits (n = 29), as a minimum of two repeated growth measurements were required to fit the latent class growth trajectory models. The KPNC Institutional Review Board approved this study (CN‐08EGund‐01‐H_29 and CN‐04EGund‐03‐H_77). All participants provided written informed consent for all study activities and assessments in the mothers and their infants [20].
2.2. Measures
2.2.1. Infant Anthropometric Measurements
Pregnancy course and perinatal outcome data were extracted from the KPNC electronic health records (EHR) system, including weight (birth, hospital scale), length (measuring board), and head circumference (measuring tape) measured at birth by KPNC clinical hospital staff. At each of the three in‐person research visits—at infant ages 6–9 weeks, 6 months, and 12 months—trained research staff obtained measurements using calibrated research equipment following the WHO Multi‐center Growth Reference Study standardized procedures [25]. Suprailiac skinfold, not described in the WHO, was measured using the following protocol: “Measure along the midaxillary line (center of body) just above the iliac crest (top of hip bone) and make a mark. With thumb on mark and forefinger anterior to mark, “pinch” the skin that follows the natural fold which will follow a line of approximately from the suprailiac to the umbilicus (bellybutton), an angle of approximately 30 degrees. Place callipers perpendicular to skinfold and take the readings after approximately 5 s. Measure to the nearest 0.1–0.5 mm in triplicate, and record measurements and any modifications to the protocol on the infant anthropometry form.” Weight (infant scale, Tanita, Model BD590, Illinois, USA), length (supine using infantometer, Seca North America, Model 417, Ca, USA), and head circumference (research non‐stretchable measuring tape) were measured in triplicate. Skinfold thicknesses (triceps, suprailiac, and subscapular), were also measured in triplicate, with Holtain callipers (Seritex Inc. 1, NJ, USA). Measures were averaged to derive the value used in analysis. Z‐scores were calculated for age and sex groups based on the WHO child growth standards (http://www.who.int/childgroth/en/) as the referent population [26].
2.2.2. Lactation Intensity Ratio
Breastfeeding intensity and duration were collected using a diary from birth to 6 weeks, by surveys filled out at 6–9 weeks, 6 months, and 12 months research visits, and via monthly mailed surveys [20]. (Figure S2) The lactation intensity ratio (LIR), a quantitative measure of breastfeeding intensity, was calculated monthly (score of 0 to 1) from month 1 to 12 based on the total number of breastfeeding occurrences divided by the sum of total number of breastfeeding plus formula feedings in the last 7 days [27]. Sum of LIR (sumLIR) was calculated to investigate any summative effect of breastfeeding over time compared to the individual monthly LIR. The sumLIR was the total for the current month plus each preceding month—for example, sumLIR 2 is LIR from months 1 and 2 (0 to 2) added while sumLIR 3 is equal to the addition of LIR from months 1 to 3 (0 to 3). The sumLIR12 is for the entire first year of life (0 to 12). Breastfeeding duration was calculated to the nearest month based on responses from questions from in‐person research visits and the monthly surveys on the date of breastfeeding cessation.
2.2.3. Infant Dietary Intake
The SWIFT Study developed the in‐person research visit surveys, administered by trained research assistants, and the monthly mailed surveys to collect data on infant health and dietary intake [22]. The in‐person surveys asked parents to report the number of breastfeeds and the number of formula feeds and total feeds per day, as well as the amount of formula fed per feeding. The surveys asked about the types of beverages given to the infants, including plain water and added sugar. They also queried parents about food and beverage intake, complementary food introduction and sugary beverages (i.e., 100% fruit juice (FJ) and sugar‐sweetened beverage (SSB) intake). Beverage intake was classified into three groups: SSB if an infant had SSB intake regardless of 100% FJ intake; FJ if an infant had 100% FJ and no SSB and None if an infant had neither SSB or FJ.
2.2.4. Covariates
Maternal delivery data at birth was collected from hospital EHR. Study surveys were completed by mothers to collect information on the Special Supplemental Nutrition Program for Women, Infants and Children (WIC) participation, reproductive and health history, intercurrent pregnancies, and lifestyle behaviours. Maternal covariates also included age, pre‐pregnancy BMI, educational attainment, gestational weight gain, parity, and GDM treatment type, gestational age at GDM diagnosis, and glucose tolerance at diagnosis from the EHR. Infant covariates include race and ethnicity, and gestational age at birth. Sugary beverage intake (fruit juice (FJ) and SSB during the first year of life) was also added as a covariate in a sensitivity analysis. Other research from this cohort has shown no significant differences between severity of prenatal glucose intolerance (3‐h 100 g OGTT z‐score sum) and breastfeeding intensity and duration measures; which is why we did not include prenatal glucose intolerance as a covariate in the models for this study [28].
2.2.5. Statistical Analysis
All analyses were conducted using R version 4.3.0 with an alpha of 0.05. Descriptive statistics and differences between those included and excluded from the analysis were calculated with t‐tests and chi‐square tests. Our primary analyses were conducted in steps. First, we fit Latent class mixed models (LCMM) using the lcmm package in R with z‐scores for weight‐for‐age (WAZ), height‐for‐age (LAZ), and weight‐for‐length (WLZ) as well as raw values for triceps, suprailiac, subscapular skinfold thickness and sum of skinfolds (summation our three measures) [29]. As two of the four measurements (birth and 6–9 weeks) were infant ages prior to 3‐months, and WHO z‐scores do not start until 3 months for skinfold measurements, we could not use z‐scores for trajectory models of skinfolds. LCMM models use individual data to identify distinct class (trajectory) patterns, then group individuals with similar patterns into a class based on the highest probability of membership. Best fit LCMM models were chosen based on models that had a minimum of 5% in each cell, with the lowest Bayesian Information Criteria (BIC). Then Average Posterior Probability of Assignment (APPA) and Odds of Correct Classification (OCC) were calculated to confirm model fit and strength. (Table S1).
After identifying the best fit LCMM models, we applied multinomial logistic regression (mlogit), using the mlogit package in R [30], to examine associations between the initial LIR and sumLIRs with LCMM trajectory class membership for each growth measure, adjusting for covariates including pre‐pregnancy BMI, gestational age at diagnosis of GDM, GDM treatment type (oral medication, insulin or diet alone), WIC participation, racial and ethnic groups, infant sex, maternal education, and parity. The reference group in each z‐score model was the group whose growth was closest to the median z‐score of zero, while for skinfold thickness, reference groups were chosen based on the largest group number.
A series of sensitivity analyses were conducted. A second set of mlogit models was fit, with sugary beverages (SSB and FJ) added as covariates, to assess whether the effects of breastfeeding on growth were influenced by sugary beverage intake. Because skinfold thickness z‐scores are not available before 3 months of age, thus preventing use of LCMM, we used linear regression models to evaluate the association between sumLIR and the change in z‐score for skinfolds (triceps and subscapular) from 6 months to 12 months.
3. Results
Of the 464 infants enrolled in the SWIFT Offspring Study, 435 were included in the analytic sample and 29 were excluded from this analysis because they only had one anthropometric measurement (Figure 1). Maternal and infant demographic characteristics are shown in Table 1. Most baseline characteristics were similar between those included versus excluded from the analysis. The only difference at baseline between those included and excluded from the analysis was beverage type; however, 41% of participants excluded from the sample had missing data for beverages, thus it is unknown if the difference is due to missing data versus differences in population. The sample was racially and ethnically diverse, with two‐thirds of mothers having managed their GDM through diet modification (65.7%).
FIGURE 1.
Flow diagram for analytic sample.
TABLE 1.
Maternal and infant characteristics at baseline, SWIFT Offspring a .
| Characteristic | SWIFT Offspring N = 464 | Analytic sample N = 435 | Excluded N = 29 | p |
|---|---|---|---|---|
| Maternal | ||||
| Age at delivery (years) | 33.1 ± 4.7 | 33.2 ± 4.7 | 31.8 ± 3.9 | 0.10 |
| Education | ||||
| High School or less | 118 (25.4) | 106 (24.4) | 12 (41.4) | |
| Some college | 134 (28.9) | 126 (28.9) | 8 (27.6) | 0.10 |
| 4 years of college or more | 212 (45.7) | 203 (46.7) | 9 (31.0) | |
| Race/ethnicity | ||||
| Non‐Hispanic White | 130 (28.0) | 124 (28.5) | 6 (20.7) | 0.51 |
| Non‐Hispanic Black | 39 (8.4) | 36 (8.3) | 3 (10.3) | |
| Hispanic | 166 (35.8) | 157 (36.1) | 9 (31.1) | |
| Asian | 120 (25.8) | 109 (25.0) | 11 (37.9) | |
| Multi‐racial/Other | 9 (1.9) | 9 (2.1) | 0 (0) | |
| Pre‐pregnancy BMI (kg/m2) | 30.3 ± 7.3 | 30.3 ± 7.5 | 30.8 ± 6.0 | 0.72 |
| WIC recipient | 130 (28.0) | 123 (28.3) | 7 (24.1) | 0.63 |
| GA at GDM diagnosis (weeks) | 25.1 ± 7.7 | 25.1 ± 7.8 | 25.7 ± 7.3 | 0.67 |
| Gestational weight gain, kg | 10.1 ± 7.0 | 10.2 ± 6.9 | 9.1 ± 7.7 | 0.40 |
| Prenatal 3‐h OGTT z‐score | 0.1 ± 2.8 | 0.8 ± 2.9 | 0.0 ± 2.8 | 0.15 |
| GDM Treatment | 0.46 | |||
| Diet modification only | 306 (65.9) | 286 (65.7) | 20 (69.0) | |
| Oral medications b | 136 (29.4) | 127 (29.2) | 9 (31.0) | |
| Insulin | 22 (4.7) | 22 (5.1) | 0 (0) | |
| First pregnancy | 153 (33.0) | 144 (33.1) | 9 (31.0) | 0.82 |
| Infant | ||||
| Female Sex | 210 (45.2) | 197 (45.3) | 13 (44.8) | 0.96 |
| GA at birth (weeks) | 39.0 ± 1.1 | 39.0 ± 1.1 | 39.1 ± 1.4 | 0.71 |
| Breastfeeding duration (months) | 9.1 ± 7.7 | 9.4 ± 7.2 | 9.1 ± 8.8 | 0.83 |
| Other Beverage Intake (Birth to 1 year) c | < 0.001 | |||
| None | 148 (31.9) | 143 (32.9) | 5 (17.2) | |
| Fruit juice (100% natural) | 219 (47.2) | 214 (49.2) | 5 (17.2) | |
| Sugar sweetened beverage | 81 (17.5) | 74 (17.0) | 7 (24.1) | |
Abbreviations: BMI, body max index (kg/m2); GA, gestational age; GDM, gestational diabetes; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children.
The chi‐square test was used for categorial variables; t‐test for continuous variables.
99% of patients on glyburide.
Sixteen participants had missing data: 4 sample; 12 from those excluded; Fisher exact test used for small sample size.
Infant growth and feeding characteristics across the study are displayed in Table 2. Average lactation intensity ratio for each month decreased from birth to 12 months, while sum lactation intensity ratio (sumLIR, max 12) increased slowly and had a standard deviation that increased from 0.60 at birth to 4.57 by 12 months (Table 2). After conducting the LCMM analyses, the best fit models identified 3 to 5 growth trajectory classes with distinct patterns of longitudinal changes, depending on the growth measure. Referent classes for z‐scores were chosen based on a pattern of catch‐down growth with the class nearest the median, which is a z‐score of 0. For raw anthropometric measures (e.g., skinfolds), the referent classes were chosen based on the class with the most infants.
TABLE 2.
Selected infant dietary intake and anthropometric measurements in analytic sample (n = 435).
| Birth (n = 435) | 6–9 weeks (n = 431) | 6 months (n = 371) | 12 months (n = 338) | |
|---|---|---|---|---|
| Age (months) | 0 | 1.6 ± 0.3 | 6.2 ± 0.7 | 12.2 ± 0.5 |
| Weight (g) | 3425.6 ± 500 | 4934.7 ± 679 | 7711.9 ± 947 | 9882.6 ± 1472 |
| Length (cm) | 50.7 ± 2.3 | 56.1 ± 2.5 | 66.4 ± 2.7 | 75.6 ± 4.5 |
| WAZ | 0.25 ± 1.0 | −0.16 ± 0.9 | −0.05 ± 0.9 | 0.19 ± 1.0 |
| LAZ | 0.61 ± 1.2 | −0.23 ± 1.7 | −0.28 ± 1.1 | −0.26 ± 1.1 |
| WLZ | −0.32 ± 1.3 | 0.12 ± 1.2 | 0.25 ± 1.0 | 0.43 ± 1.0 |
| BMIZ | −0.10 ± 1.1 | −0.20 ± 1.0 | 0.15 ± 1.0 | 0.38 ± 1.0 |
| Skinfolds (mm) | ||||
| Triceps | NA | 9.8 ± 2.2 | 10.7 ± 2.4 | 10.5 ± 2.5 |
| Subscapular | NA | 8.4 ± 2.2 | 8.2 ± 2.0 | 7.7 ± 2.0 |
| Sum of skinfolds | NA | 28.2 ± 7.1 | 29.1 ± 7.9 | 26.7 ± 7.8 |
| Lactation Intensity Ratio | 0.80 ± 0.3 | 0.74 ± 0.4 | 0.51 ± 0.4 | 0.31 ± 0.4 |
| Sum of LIR | NA | 1.55 ± 0.6 | 3.89 ± 2.2 | 6.25 ± 4.6 |
Abbreviations: BMIZ, body mass index‐for‐age z‐score; LAZ, length‐for‐age z‐score; LIR, lactation intensity ratio; WAZ, weight‐for‐age z‐score; WLZ, weight‐for‐length z‐score.
For WLZ, the best fit model identified 5 trajectory classes (Figure 2a). Two classes (class 1 and 2) showed catch down growth, characterized by a decrease in z‐score until 6 months, followed by plateau until 10 months. At the 12‐month visit, there was a slow steady WLZ gain; however, z‐score changes were less than 1SD from birth to 1 year, indicating appropriate changes in weight for length The other three classes (3–5) did not experience catch‐down growth. Strikingly, class 3 showed a z‐score increase of more than 2 SD from birth to 6 months. Classes 4 and 5 increased in WLZ from birth to 6–9 weeks, then plateaued. In our multinomial logistic regression models, higher cumulative breastfeeding intensity summed monthly was associated with lower odds of membership in class 3 and 4 for all months from 1 to 12 (p < 0.01 for class 3, p < 0.05 class 4) and class 5 for the first 5 months but not thereafter compared to class 2. By month 12, a higher sumLIR was associated with 16% (OR 0.84 [95% CI 0.75, 0.95]) lower odds of belonging to class 3 and a 7% (OR 0.93 [0.87, 0.99]) lower odds of belonging to class 4, both classes with less optimal growth patterns compared to class 2 (Table 3).
FIGURE 2.
Latent Class Mixed Models (LCMM) showing class trajectory classes illustrating data driven concomitant patterns of change with 95% confidence interval bands for (a) weight‐for‐length z‐score (WLZ); (b) length‐for‐age z‐score (LAZ); (c) weight‐for‐age z‐score (WAZ); (d) sum of skinfold thickness, mm; (e) subscapular skinfold thickness, mm; (f) triceps skinfold thickness, mm. Percentages are posterior probabilities for class membership reflecting the proportion of the sample within each class.
TABLE 3.
Multinomial logistic regression associations between sum LIR and LCMM class, adjusted models.
| Sum of lactation intensity ratio (sumLIR) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Class | LIR 1 | Sum 2 | Sum 3 | Sum 4 | Sum 5 | Sum 6 | Sum 7 | Sum 8 | Sum 9 | Sum 10 | Sum 11 | Sum 12 |
| Weight‐for‐length z‐score | ||||||||||||
| 1 |
0.51 0.11–2.40 |
0.79 0.39, 1.58 |
0.87 0.56–1.34 |
0.92 0.67–1.25 |
0.92 0.73–1.17 |
0.93 0.77–1.13 |
0.95 0.81–1.11 |
0.96 0.84–1.10 |
0.97 0.86–1.09 |
0.97 0.87–1.08 |
0.97 0.88–1.07 |
0.98 0.89–1.06 |
| 2 | ref | ref | ref | ref | ref | ref | ref | ref | ref | ref | ref | ref |
| 3 |
0.11** 0.02–0.54 |
0.39* 0.18–0.82 |
0.52* 0.32–0.83 |
0.60** 0.43–0.85 |
0.66** 0.51–0.87 |
0.72** 0.57–0.90 |
0.76** 0.63–0.92 |
0.78** 0.66–0.92 |
0.80** 0.69–0.93 |
0.82** 0.72–0.94 |
0.83** 0.73–0.94 |
0.84** 0.75–0.95 |
| 4 |
0.22** 0.07, 0.68 |
0.51** 0.31–0.85 |
0.66* 0.48–0.90 |
0.74** 0.59–0.92 |
0.79** 0.66–0.93 |
0.83** 0.72–0.95 |
0.85** 0.76–0.96 |
0.88** 0.80–0.97 |
0.89** 0.82–0.97 |
0.91* 0.84–0.98 |
0.92* 0.86–0.98 |
0.93* 0.87–0.99 |
| 5 |
0.17** 0.05, 0.60 |
0.49* 0.29–0.86 |
0.66* 0.47–0.93 |
0.75* 0.59–0.96 |
0.81* 0.67–0.98 |
0.85* 0.73–0.99 |
0.88 0.78–1.00 |
0.91 0.81–1.01 |
0.92 0.84–1.01 |
0.94 0.86–1.02 |
0.94 0.88–1.02 |
0.95 0.89–1.02 |
| Sum of Skinfolds, mm | ||||||||||||
| 1 |
0.99 0.40–2.44 |
1.16 0.77–1.75 |
1.07 0.83–1.39 |
1.06 0.88–1.28 |
1.03 0.89–1.20 |
1.02 0.91–1.15 |
1.02 0.92–1.12 |
1.01 0.93–1.10 |
1.01 0.94–1.09 |
1.01 0.94–1.08 |
1.00 0.95–1.07 |
1.00 0.95–1.06 |
| 2 |
0.50 0.23–1.10 |
0.82 0.57–1.17 |
0.87 0.70–1.10 |
0.91 0.77–1.07 |
0.93 0.82–1.06 |
0.95 0.85–1.05 |
0.96 0.88–1.05 |
0.96 0.89–1.04 |
0.97 0.91–1.04 |
0.97 0.92–1.03 |
0.97 0.92–1.03 |
0.98 0.93–1.03 |
| 3 | ref | ref | ref | ref | ref | ref | ref | ref | ref | ref | ref | ref |
| Weight for age z‐score | ||||||||||||
| 1 |
0.38* 0.16–0.94 |
0.67 0.44–1.04 |
0.78 0.59–1.02 |
0.85 0.69–1.03 |
0.88 0.75–1.02 |
0.90 0.79–1.02 |
0.92 0.82–1.02 |
0.92 0.84–1.02 |
0.93 0.85–1.01 |
0.93 0.87–1.01 |
0.94 0.87–1.00 |
0.94 0.88–1.00 |
| 2 | ref | ref | ref | ref | ref | ref | ref | ref | ref | ref | ref | ref |
| 3 |
0.65 0.21–1.99 |
0.80 0.47–1.34 |
0.80 0.58–1.10 |
0.83 0.65–1.05 |
0.84 0.70–1.01 |
0.86 0.74–1.00 |
0.87* 0.77–0.99 |
0.88* 0.79–0.99 |
0.89* 0.81–0.99 |
0.90* 0.82–0.98 |
0.91* 0.83–0.98 |
0.91* 0.84–0.98 |
| 4 |
0.80 0.34–1.86 |
0.88 0.60–1.30 |
0.92 0.72–1.17 |
0.94 0.79–1.12 |
0.95 0.83–1.09 |
0.96 0.86–1.07 |
0.97 0.88–1.06 |
0.97 0.90–1.05 |
0.97 0.91–1.04 |
0.97 0.91–1.04 |
0.98 0.92–1.03 |
0.98 0.93–1.03 |
Note: Models were adjusted for gestational age at diagnosis of diabetes, GDM treatment type (medication, insulin, or diet alone), Supplemental Program for Women, Infant, and Children (WIC) participation, race/ethnicity, pre‐pregnancy BMI, infant sex, maternal education, and parity; *p < 0.05, **p < 0.01.
For WAZ, four trajectory classes were identified. (Figure 2c). All four classes experienced some catch‐down growth. However, Class 3 experienced the smallest catch‐down and an increase in z‐score 1.5 SD from birth to 1 year. Two Classes (1 and 4) had catch‐down growth with slight increase in z‐score while Class 2 plateaued until 1‐year. Higher sumLIR was associated with a lower odds of being in Class 3 from 7 months (OR 0.87 [0.77–0.99]) to 12 months (OR 0.91 [0.84–0.99]) (Table 3) with a similar pattern for monthly LIR (Table S2).
For sum of skinfolds (SS), three trajectory classes were identified. Class 3 stayed steady with measurements within 5 mm of each other across infancy (between 21 and 24 mm), whereas Class 2 had a steady decline with skinfolds staying within 5 mm from beginning (30 mm) to end (25 mm); while Class 1 initially had an increase in skinfold thickness from birth to 6 months followed by decline and ending around the same measurement as birth (Figure 2d). LIR, monthly or summed, was not associated with sum of skinfolds. (Table 3, Table S3).
Unadjusted models for all regression analyses can be found in Table S4 (monthly LIR) and Table S5 (sumLIR) in the Supporting Information. The odds ratios did not change appreciably when sugary beverages were added as a covariate; therefore, sugary beverages were not included in any model.
LIR was associated with some changes in skinfold z‐scores from 6 to 12 months (Table 4). SumLIR at 12 months was inversely associated with z‐score change −0.03 (95% CI, −0.07 to −0.003) for SSF, indicating that for every 1‐unit increase in lactation intensity, the sum of skinfold thickness was reduced. For TSF, each unit higher of sumLIR was associated with a 0.05 (95% CI, −0.08 to −0.01) unit reduction in TSF z‐score from 6 to 12 months.
TABLE 4.
Linear regression evaluating association between each unit increase in sum LIR at 12 months and change in z‐score from 6 to 12 months for selected skinfolds (n = 318).
| z‐score | sumLIR 12 | |
|---|---|---|
| β (95% CI) | ||
| Subscapular Skinfold | −0.03 (−0.07 to −0.003)* | |
| Triceps Skinfold | −0.05 (−0.08 to −0.01)** |
Note: Models were adjusted for gestational age at diagnosis of diabetes, GDM treatment type (medication, insulin, or diet alone), Special Supplemental Program for Women Infant and Children (WIC) participation, race/ethnicity, pre‐pregnancy BMI, infant sex, maternal education, and parity; *p < 0.05, **p < 0.01.
4. Discussion
In our large prospective study of 435 infants exposed to GDM in utero and followed during the first year of life, higher breastfeeding intensity score, measured both monthly and cumulatively for each month across infancy from birth to 12 months, was positively associated with a steadier rate of slower growth patterns for WAZ and WLZ but was not associated with adipose tissue accrual over time.
Three previous studies used latent class mixed models to evaluate growth trajectories in infants exposed to GDM, but none of these considered infant feeding practices [1, 31, 32]. One study showed that GDM‐exposed infants had lower body fat at birth, but then interestingly showed lower adipose tissue accrual patterns from birth to 12 months, indicating a downshift in growth; however, milk feeding modality was not assessed [1]. Two others evaluated maternal factors and infant growth associations, but not infant milk‐feeding type or other dietary intake [31, 32]. Our study is the first (to our knowledge) to evaluate a quantitative breastfeeding intensity measure, assessed monthly, on associations with growth trajectories developed using advanced growth modelling techniques. Recently published research in this population has demonstrated infant growth related to milk type independent of the prenatal environment [33] and we adjust for prenatal exposures, including GDM treatment type and timing of diagnosis during gestation, in our analyses.
Our results are similar to findings on changes in growth rates over time with breastfeeding associated with slower overall average growth in the first year. In 348 exclusively breastfed infants both exposed and unexposed to GDM, Nagel et al. found similar rates of growth [34]. Faith et al. found in 382 infants that intensive breastfeeding as well as breastfeeding duration up to 12 months were associated with lower odds of obesity [35]. In SWIFT Offspring, our team previously found that changes in WLZ from birth to 1 year were lower on average with consistent high‐intensity breastfeeding versus consistent high formula feeding [20]. These current SWIFT Offspring findings add to this growing body of literature, demonstrating that it is not just the presence of any breastfeeding duration but rather the amount or intensity of breastmilk feeding that is important in infants exposed in utero to GDM. Both the monthly LIR and sumLIR of breastfeeding over the first year were associated with a reduced likelihood of being in the high weight gain classes for WLZ and WAZ indicating that even if not exclusive, breastfeeding is protective in early life to slow weight trajectories among GDM exposed infants. Despite studies showing that breastmilk from mothers with GDM may have more unfavourable components [36, 37], our results and those from other studies demonstrate that breastfeeding should still be considered as a beneficial method of infant feeding.
Few have reported skinfold thickness changes across the first year of life in GDM and none have used latent class trajectory methods [38, 39]. In contrast to our results, Cwiek et al. found breastfeeding (defined as breastfeeding, mixed feeding, and formula feeding) was associated with lower subscapular skinfold in boys, but did not quantify breastfeeding change across time, studied only 42 infants exposed in utero to GDM, and did not use trajectory class methods [38]. In our analyses, we found no associations between monthly LIR or sumLIR across infancy with subcutaneous adipose tissue raw measures using triceps and subscapular trajectory groups from LCMM models. However, when we compared the mean change in the triceps and subscapular z‐score measurements in linear regression models in a secondary analysis we found that each unit increase in sumLIR was associated with a decrease in skinfold‐for‐age z‐score from 6 to 12 months; suggesting that effects of breastfeeding intensity on skinfolds are more apparent in second half of infancy, and/or that use of the WHO growth reference for skinfolds to examine predictors of skinfolds over time may offer more insight about early life exposures compared to raw skinfold thickness values. While the beta coefficients are small in our skinfold thickness linear regression models, they represent a 1‐unit change in LIR, which is not clinically significant for 1‐sumLIR change but can be more meaningful for those with higher sumLIR. For example, for the mean sumLIR value of 6.2, this represents an effect size of 0.19 lower subscapular skinfold z‐score and 0.31 lower triceps skinfold z‐score, and at the highest level for sumLIR of 12, this represents an estimated effect size of 0.36 lower subscapular skinfold z‐score and 0.6 mm lower triceps skinfold z‐score, the latter of which approaches crossing a percentile on the WHO skinfold growth chart (i.e., a z‐score change of 0.67), which is a clinicially meangingful change.
5. Limitations
Our study is not without limitations. First, breastfeeding was self‐reported, so there may be some reporting bias; however, to minimize bias, data were collected prospectively over several days during each of the 12 months. Second, skinfold z‐scores are not available until 3 months of age, so raw scores were used for LCMM. Third, this study included only mother‐infant dyads exposed to GDM but assessed detailed data on GDM severity. Notably, there is a plethora of research on breastfeeding among infants without in utero GDM exposure. Maternal prenatal diet was not included in these models, but a recent paper published on the larger cohort demonstrated that maternal diet, specifically prescribed as the cornerstone of GDM treatment, was not statistically significantly associated with risk of obesity in SWIFT children [33].
Strengths of this study include our repeated longitudinal anthropometric measurements, coupled with monthly, prospectively collected breastfeeding intensity and duration data, using quantitative methods [27]. The data also include a complete range of maternal prenatal laboratory testing, as well as data on GDM diagnosis and management, gestational weight gain, and perinatal outcomes for mothers and newborns. Finally, this GDM population includes diverse infants, with a high retention rate (88% at 1 year).
6. Conclusions
In our cohort, higher lactation intensity was associated with more optimal WLZ trajectories, and with a decrease in subscapular and tricipes skinfold thickness z‐scores from 6 to 12 months. These findings are likely generalizable to GDM offspring in the U.S., given the SWIFT cohort is representative of the diversity of U.S. births to women with GDM [23]. Although exclusive breastfeeding for 6 months is the optimal feeding choice, it may not be feasible for all mothers. In those instances when mothers cannot exclusively breastfeed, mixed feedings with as much breastfeeding should be encouraged, especially in infants exposed to GDM in utero as a higher intensity of breastfeeding may support healthier growth and body composition changes during infancy. Prenatal education and breastfeeding support should be made more available to postpartum women with obesity and GDM, who often encounter greater barriers to sustaining breastfeeding due to higher perinatal complication rates.
Author Contributions
The authors thank all physicians, clinical staff, research staff, and the SWIFT Offspring participants for their commitment and important contributions to this study. Dr. Gunderson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: Erica P. Gunderson, Elizabeth M. Widen, Saralyn F. Foster. Drafting of the manuscript: Saralyn F. Foster, Erica P. Gunderson, Elizabeth M. Widen. Critical revision of the manuscript for important intellection content: Saralyn F. Foster, Elizabeth M. Widen, Joan C. Lo, Louise C. Greenspan, Myles S. Faith, Baiyang Sun, Alexis King, Marcela R. Abrego, Daniel Powers, Erica P. Gunderson. Statistical analysis: Erica P. Gunderson, Daniel Powers, Elizabeth M. Widen, Baiyang Sun, Saralyn F. Foster. Obtained funding: Erica P. Gunderson. Administrative, technical or material support: Joan C. Lo, Louise C. Greenspan, Alexis King. Supervision: Elizabeth M. Widen, Erica P. Gunderson.
Funding
The SWIFT Study and the SWIFT Offspring Study were supported by grants to Dr. Erica P. Gunderson from the NICHD (R01 HD050625), The American Diabetes Association (Clinical Research Award), and the NIDDK (R21 DK103171). Additional support for this manuscript was supported by a grant to Dr. Elizabeth Widen NICHD (R00HD086304).
Disclosure
The authors have nothing to report.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table S1: Model Fit Statistics by Anthropometric Measure
Table S2: Multinomial logistic regression associations between LIR and LCMM class, adjusted models
Table S3: Multinomial logistic regression associations between sumLIR and LCMM class for selected anthropometrics, adjusted models
Table S4: Multinomial logistic regression associations between LIR and LCMM class, unadjusted modelsa
Table S5: Multinomial logistic regression associations between sumLIR and LCMM class, unadjusted modelsa
Figure S1: Scatterplots of z‐score change from 6 month to 12 months for subscapular and triceps skinfolds. sumLIR 12 is the cumulative sum of the lactation intensity ratio (LIR) from birth to 12 months, scale 0–12, where higher number indicates more breastfeeding, LIR 12 = breastfeeding for 12 months, LIR 0 = no breastfeeding.
Figure S2: Research Surveys developed by the Study of Women, Infant Feeding and Type 2 Diabetes After Gestational Diabetes: THE SWIFT STUDY.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Rickman R. R., Widen E. M., Lane C. E., et al., “Infant Body Composition Trajectories Differ by in Utero Exposure to Gestational Diabetes Mellitus: A Prospective Cohort From Birth to 12 Months,” American Journal of Clinical Nutrition 121, no. 1 (2025): 40–49, 10.1016/j.ajcnut.2024.10.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Logan K. M., Gale C., Hyde M. J., Santhakumaran S., and Modi N., “Diabetes in Pregnancy and Infant Adiposity: Systematic Review and Meta‐Analysis,” Archives of Disease in Childhood. Fetal and Neonatal Edition 102, no. 1 (2017): F65–f72, 10.1136/archdischild-2015-309750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Mitanchez D., Jacqueminet S., Lebbah S., Dommergues M., Hajage D., and Ciangura C., “Relative Contribution of Gestational Weight Gain, Gestational Diabetes, and Maternal Obesity to Neonatal Fat Mass,” Nutrients 12, no. 11 (2020): 3434, 10.3390/nu12113434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Benhalima K., De Landtsheer A., Van Crombrugge P., et al., “Predictors of Neonatal Adiposity and Associations by Fetal Sex in Women With Gestational Diabetes Mellitus and Normal Glucose‐Tolerant Women,” Acta Diabetologica 58, no. 3 (2021): 341–354, 10.1007/s00592-020-01619-0. [DOI] [PubMed] [Google Scholar]
- 5. Andersson‐Hall U. K., Järvinen E. A. J., Bosaeus M. H., et al., “Maternal Obesity and Gestational Diabetes Mellitus Affect Body Composition Through Infancy: The PONCH Study,” Pediatric Research 85, no. 3 (2019): 369–377, 10.1038/s41390-018-0248-9. [DOI] [PubMed] [Google Scholar]
- 6. Leth‐Møller M., Kampmann U., Hede S., Ovesen P. G., Hulman A., and Knorr S., “Breastfeeding and Infant Growth in Relation to Childhood Overweight ‐ A Longitudinal Cohort Study,” American Journal of Clinical Nutrition 121 (2025): 835–842, 10.1016/j.ajcnut.2025.01.020. [DOI] [PubMed] [Google Scholar]
- 7. Cao W., Li N., Zhang R., et al., “Interactive Effects of Gestational Diabetes and High Pre‐Pregnancy Body Mass Index on Adverse Growth Patterns of Offspring,” Diabetes/Metabolism Research and Reviews 40, no. 3 (2024): e3759, 10.1002/dmrr.3759. [DOI] [PubMed] [Google Scholar]
- 8. Lai T., Mao A., Yang L., et al., “Association of Maternal Metabolic Risk Factors With Offspring Body Mass Index (BMI) Trajectories in Early Childhood: A Retrospective Cohort Study,” BMJ Open 15, no. 3 (2025): e088641, 10.1136/bmjopen-2024-088641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Cho S., Rifas‐Shiman S. L., Harnois‐Leblanc S., Aris I. M., Oken E., and Hivert M. F., “Associations of Abnormal Maternal Glucose Regulation in Pregnancy With Offspring Adiposity, Insulin Resistance, and Adipokine Markers During Childhood and Adolescence,” Journal of Pediatrics 272 (2024): 114100, 10.1016/j.jpeds.2024.114100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Li W., Wang L., Liu H., et al., “Maternal Gestational Diabetes and Childhood Adiposity Risk From 6 to 8 Years of Age,” International Journal of Obesity 48, no. 3 (2024): 414–422, 10.1038/s41366-023-01441-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Huvinen E., Tuomaala A. K., Bergman P. H., et al., “Ascending Growth Is Associated With Offspring Adiposity in Pregnancies Complicated With Obesity or Gestational Diabetes,” Journal of Clinical Endocrinology and Metabolism 106, no. 5 (2021): e1993–e2004, 10.1210/clinem/dgaa979. [DOI] [PubMed] [Google Scholar]
- 12. Otter G., Davis D., Kurz E., et al., “Promoting Breastfeeding in Women With Gestational Diabetes Mellitus in High‐Income Settings: An Integrative Review,” International Breastfeeding Journal 19, no. 1 (2024): 4, 10.1186/s13006-023-00603-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Dewey K. G., Heinig M. J., Nommsen L. A., Peerson J. M., and Lönnerdal B., “Growth of Breast‐Fed and Formula‐Fed Infants From 0 to 18 Months: The DARLING Study,” Pediatrics 89, no. 6 Pt 1 (1992): 1035–1041. [PubMed] [Google Scholar]
- 14. Santos I. S., Costa C. S., Hills A. P., et al., “Infant Body Composition at 6 and 24 Months: What Are the Driving Factors?,” European Journal of Clinical Nutrition 78, no. 11 (2024): 928–935, 10.1038/s41430-023-01321-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Camier A., Cissé A. H., Heude B., et al., “Infant Feeding Practices and Body Mass Index up to 7.5 Years in the French Nationwide ELFE Study,” Pediatric Obesity 19, no. 7 (2024): e13121, 10.1111/ijpo.13121. [DOI] [PubMed] [Google Scholar]
- 16. Beyerlein A., Toschke A. M., and von Kries R., “Breastfeeding and Childhood Obesity: Shift of the Entire BMI Distribution or Only the Upper Parts?,” Obesity (Silver Spring) 16, no. 12 (2008): 2730–2733, 10.1038/oby.2008.432. [DOI] [PubMed] [Google Scholar]
- 17. Longmore D. K., Titmuss A., Barr E., et al., “Breastfeeding and Infant Growth in Offspring of Mothers With Hyperglycaemia in Pregnancy: The Pregnancy and Neonatal Diabetes Outcomes in Remote Australia Study,” Pediatric Obesity 17, no. 6 (2022): e12891, 10.1111/ijpo.12891. [DOI] [PubMed] [Google Scholar]
- 18. Vandyousefi S., Whaley S. E., Widen E. M., et al., “Association of Breastfeeding and Early Exposure to Sugar‐Sweetened Beverages With Obesity Prevalence in Offspring Born to Mothers With and Without Gestational Diabetes Mellitus,” Pediatric Obesity 14, no. 12 (2019): e12569, 10.1111/ijpo.12569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Filardi T., Bleve E., Gorini S., Caprio M., and Morano S., “Is Breastfeeding an Effective Approach to Reduce Metabolic Risk After GDM in Mothers and Infants?,” Journal of Clinical Medicine 14, no. 9 (2025): 3065, 10.3390/jcm14093065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Gunderson E. P., Greenspan L. C., Faith M. S., Hurston S. R., C. P. Quesenberry, Jr. , and Investigators SOS , “Breastfeeding and Growth During Infancy Among Offspring of Mothers With Gestational Diabetes Mellitus: A Prospective Cohort Study,” Pediatric Obesity 13, no. 8 (2018): 492–504, 10.1111/ijpo.12277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Vandyousefi S., Davis J. N., and Gunderson E. P., “Association of Infant Diet With Subsequent Obesity at 2‐5 Years Among Children Exposed to Gestational Diabetes: The SWIFT Study,” Diabetologia 64, no. 5 (2021): 1121–1132, 10.1007/s00125-020-05379-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Gunderson E. P., Matias S. L., Hurston S. R., et al., “Study of Women, Infant Feeding, and Type 2 Diabetes Mellitus After GDM Pregnancy (SWIFT), A Prospective Cohort Study: Methodology and Design,” BMC Public Health 11 (2011): 952, 10.1186/1471-2458-11-952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Gunderson E. P., Hurston S. R., Dewey K. G., et al., “The Study of Women, Infant Feeding and Type 2 Diabetes After GDM Pregnancy and Growth of Their Offspring (SWIFT Offspring Study): Prospective Design, Methodology and Baseline Characteristics,” BMC Pregnancy and Childbirth 15 (2015): 150, 10.1186/s12884-015-0587-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus , “Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus,” Diabetes Care 23, no. Suppl 1 (2000): S4–S19. [PubMed] [Google Scholar]
- 25. de Onis M., Onyango A. W., Van den Broeck J., Chumlea W. C., and Martorell R., “Measurement and Standardization Protocols for Anthropometry Used in the Construction of a New International Growth Reference,” Food and Nutrition Bulletin 25, no. 1 Suppl (2004): S27–S36, 10.1177/15648265040251s104. [DOI] [PubMed] [Google Scholar]
- 26. WHO Multicentre Growth Reference Study Group , “WHO Child Growth Standards Based on Length/Height, Weight and Age,” Acta Paediatrica (Oslo, Norway: 1992) 95, no. 450 (2006): 76–85. [DOI] [PubMed] [Google Scholar]
- 27. Piper S. and Parks P. L., “Use of an Intensity Ratio to Describe Breastfeeding Exclusivity in a National Sample,” Journal of Human Lactation 17, no. 3 (2001): 227–232, 10.1177/089033440101700306. [DOI] [PubMed] [Google Scholar]
- 28. Gunderson E. P., Hurston S. R., Ning X., et al., “Lactation and Progression to Type 2 Diabetes Mellitus After Gestational Diabetes Mellitus: A Prospective Cohort Study,” Annals of Internal Medicine 163, no. 12 (2015): 889–898, 10.7326/m15-0807. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Proust‐Lima C. P. V., Diakite A., and Liquet B., “lcmm: Extended Mixed Models Using Latent Classes and Latent Processes. R package version: 2.2.1”, https://cran.r‐project.org/package=lcmm.
- 30. Croissant Y., “Estimation of Random Utility Models in R: The Mlogit Package,” Journal of Statistical Software 95, no. 11 (2020): 1–41. [Google Scholar]
- 31. Hu Z., Tylavsky F. A., Han J. C., et al., “Maternal Metabolic Factors During Pregnancy Predict Early Childhood Growth Trajectories and Obesity Risk: The CANDLE Study,” International Journal of Obesity 43, no. 10 (2019): 1914–1922, 10.1038/s41366-019-0326-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Thøgersen M., Nielsen K. K., Wibaek R., et al., “Body Mass Index Trajectories in Children Exposed to Gestational Diabetes in Utero: A Nationwide Register‐Based Study,” Journal of Clinical Endocrinology and Metabolism 108, no. 12 (2023): e1551–e1559, 10.1210/clinem/dgad384. [DOI] [PubMed] [Google Scholar]
- 33. Sun B., Lo J. C., Greenspan L. C., et al., “Fetal Exposure to Gestational Diabetes Severity and Postnatal Infant Feeding in the First Year of Life Associated With Preadolescent Obesity: A Prospective Cohort,” Obesity (Silver Spring) 33, no. 5 (2025): 996–1010, 10.1002/oby.24261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Nagel E. M., Peña A., Dreyfuss J. M., et al., “Gestational Diabetes, the Human Milk Metabolome, and Infant Growth and Adiposity,” JAMA Network Open 7, no. 12 (2024): e2450467, 10.1001/jamanetworkopen.2024.50467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Faith M. S., Hittner J. B., Hurston S. R., et al., “Association of Infant Temperament With Subsequent Obesity in Young Children of Mothers With Gestational Diabetes Mellitus,” JAMA Pediatrics 173, no. 5 (2019): 424–433, 10.1001/jamapediatrics.2018.5199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Geddes D. T., Gridneva Z., and Perrella S. L., “Breastfeeding After Gestational Diabetes Mellitus: Maternal, Milk and Infant Outcomes,” Current Opinion in Clinical Nutrition and Metabolic Care 28, no. 3 (2025): 257–262, 10.1097/mco.0000000000001117. [DOI] [PubMed] [Google Scholar]
- 37. Ma X., Lu Y., Huang C., et al., “Analysis of Human Milk Oligosaccharides From Women With Gestational Diabetes Mellitus,” Analytical Biochemistry 696 (2025): 115689, 10.1016/j.ab.2024.115689. [DOI] [PubMed] [Google Scholar]
- 38. Ćwiek D., Zimny M., Dawid W., et al., “Evaluation of Changes in the Anthropometric Measurements of Infants in Relation to the Type of Feeding and the Presence of Gestational Diabetes in Their Mothers: A Preliminary Study,” Journal of Clinical Medicine 14, no. 7 (2025): 2393, 10.3390/jcm14072393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Gilbert L., Quansah D. Y., Arhab A., et al., “Effect of the MySweetheart Randomized Controlled Trial on Birth, Anthropometric and Psychobehavioral Outcomes in Offspring of Women With GDM,” Frontiers in Endocrinology 14 (2023): 1148426, 10.3389/fendo.2023.1148426. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Model Fit Statistics by Anthropometric Measure
Table S2: Multinomial logistic regression associations between LIR and LCMM class, adjusted models
Table S3: Multinomial logistic regression associations between sumLIR and LCMM class for selected anthropometrics, adjusted models
Table S4: Multinomial logistic regression associations between LIR and LCMM class, unadjusted modelsa
Table S5: Multinomial logistic regression associations between sumLIR and LCMM class, unadjusted modelsa
Figure S1: Scatterplots of z‐score change from 6 month to 12 months for subscapular and triceps skinfolds. sumLIR 12 is the cumulative sum of the lactation intensity ratio (LIR) from birth to 12 months, scale 0–12, where higher number indicates more breastfeeding, LIR 12 = breastfeeding for 12 months, LIR 0 = no breastfeeding.
Figure S2: Research Surveys developed by the Study of Women, Infant Feeding and Type 2 Diabetes After Gestational Diabetes: THE SWIFT STUDY.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.