Abstract
Background/Objective:
The evidence that maternal non-nutritive sweetener (NNS) intake during pregnancy increases childhood obesity risk is conflicting. A potential reason for this is that all prior studies examined childhood body mass index (BMI) at only one timepoint and at different ages. We examined the extent to which NNS intake during pregnancy is associated with offspring BMI z-score and body fat longitudinally from birth to 18 years.
Subjects:
1683 children from Project Viva – a prospective pre-birth cohort recruited 1999–2002 in Massachusetts.
Methods:
We assessed maternal NNS intake in the first and second trimesters of pregnancy using a semi-quantitative food frequency questionnaire. Our outcomes were offspring BMI z-score, (at birth, infancy (median 6.3 months), early childhood (3.2 years), mid-childhood (7.7 years), and early adolescence (12.9 years)), sum of skinfolds (SS), fat mass index (FMI) measured by dual x-ray absorptiometry, and BMI z-score trajectory from birth to 18 years. We adjusted models for maternal pre-pregnancy BMI, age, race/ethnicity, education, parity, pre-pregnancy physical activity, smoking and paternal BMI and education.
Results:
70% of mothers were white and pre-pregnancy BMI was 24.6±5.2 kg/m2. The highest quartile of NNS intake (Q4: 0.98±0.91 servings/day) was associated with higher BMI z-score in infancy (β 0.20 units; 95% CI 0.02, 0.38), early childhood (0.21; 0.05, 0.37), mid-childhood (0.21; 0.02, 0.40), and early adolescence (0.14; −0.07, 0.35) compared with Q1 intake (Q1: 0.00±0.00 servings/day). Q4 was also associated with higher SS in early childhood (1.17mm; 0.47, 1.88), mid-childhood (2.33mm; 0.80, 3.87), and early adolescence (2.27mm; −0.06, 4.60) and higher FMI in mid-childhood (0.26kg/m2; −0.07, 0.59). Associations of maternal NNS intake with offspring BMI z-score became stronger with increasing age from 3–18 years (Pinteraction: <0.0001).
Conclusions:
Maternal NNS intake during pregnancy is associated with increased childhood BMI z-score and body fat from birth to teenage years. This is relevant given the escalating obesity epidemic, and popularity of NNS.
Introduction
The prevalence of overweight and obesity has dramatically increased in the United States, now affecting >70% of adults1. Obesity begins in early life, when 8% of infants and 26% of preschoolers are already overweight or have obesity2. This is concerning because childhood obesity strongly predicts adulthood obesity3, and comorbidities such as type 2 diabetes and cardiovascular disease4. While poor diet and reduced physical activity contribute to high obesity rates, a growing body of evidence also indicates that prenatal exposures, such as maternal high fat or high sugar5 diets, can have obesogenic effects6. There is therefore an urgent need to identify modifiable risk factors in pregnancy that could inform public health interventions and reduce the burden of obesity.
Non-nutritive sweeteners (NNSs) have become increasingly popular as an approach to reducing sugar intake and limiting associated weight gain, with >40% of adults, >25% of children, and >30% of pregnant women reporting consumption on a given day7,8. While the purpose of NNSs is to provide sweetness without calories, thereby theoretically helping to prevent obesity, recent evidence suggests a paradoxical relationship between higher NNS consumption and increased obesity risk. Proposed mechanisms for this association include impaired glucose metabolism9, altered sweet taste preference10, caloric compensation11, and dysregulation of intestinal microbiota12.
However, little is known about the effects of NNS consumption during pregnancy on the offspring’s obesity risk. In mice, NNS consumption in pregnancy induces glucose intolerance, fetal growth restriction and neonatal hypoglycemia13 – risk factors for later obesity and metabolic disease in offspring14,15. In humans, the evidence is inconsistent. While two studies found an association between maternal NNS consumption during pregnancy and increased obesity risk in offspring8,16, another saw no such association5. A potential reason for inconsistencies across studies is that they do not have repeated measures of BMI or adiposity throughout the child’s life course – instead, each study examined BMI at only one timepoint. Azad et al. examined these associations only in infancy (~1 year old)8 while Zhu et al and Gillman et al examined these associations only in mid-childhood (~7 years old)5,16. It is possible that these single timepoint analyses do not capture obesity risk across childhood, or varying associations with age. Maternal exposures often affect offspring weight differently depending on the age the outcome is assessed. For example, offspring exposed to gestational diabetes (GDM) in utero are larger at birth than unexposed offspring, similar in size in the preschool years, then have higher risk of overweight and obesity in later childhood17. For this reason, it is important to consider BMI trajectory throughout childhood, to ensure any time-varying effect of the exposure is not missed. In addition, most previous studies have also only focused on BMI, without measuring adiposity. It is also important to consider body composition alongside BMI z-scores, as BMI and body fat are only moderately correlated18 and adiposity is an independent risk factor for insulin resistance and a strong predictor of morbidity19.
Therefore, the purpose of this study was to examine the extent to which NNS intake during pregnancy is associated with offspring body mass index (BMI) z-score trajectory and body fat measures from birth to 18 years, using mother-child pairs from the pre-birth cohort Project Viva. We hypothesized that higher maternal NNS intake during pregnancy would be associated with increased body fat and childhood BMI trajectory and body fat from birth to 18 years.
Methods
Study population
Study participants were from Project Viva, a prospective pre-birth-cohort recruited between April 1999 and November 2002 from obstetric practices at Atrius Harvard Vanguard Medical Associates in Eastern Massachusetts (Clinical Trials ID: NCT02820402). Details of recruitment and retention are published elsewhere20. Of the 2128 live singleton infants, 2112 had mothers with no pre-existing diabetes and 1885 had complete food frequency questionnaire (FFQ) data during pregnancy to determine NNS intake during pregnancy. Of these, 1683 had at least 1 research-assistant-measured weight and length/height obtained at in-person research visits or ≥3 measures of weight and length/height obtained either at in-person visits or via clinical medical records (spanning birth to 18 years of age) to determine BMI trajectory (n = 1619 had in-person measured weight and length/height while n = 1570 had sufficient data for BMI trajectory). Compared with the 1683 participants in this analyses, the 445 excluded mothers were slightly younger (mean age 30.5 vs 32.2 years), had a slightly higher pre-pregnancy BMI (25.9 vs. 24.6 kg/m2), were less likely to be white (51% vs. 70%) less likely to be a college graduate (48% vs. 69%) more likely to have smoked during pregnancy (18% vs. 11%) and more likely to have gestational diabetes (9% vs. 5%). However, we noted no differences in child sex (46 vs. 49% female) or gestational age at delivery (39.2 vs. 39.5 weeks).
Mothers provided written informed consent for themselves and their children at each in-person visit. Children also provided verbal assent. The institutional review boards of Harvard Pilgrim Health Care (#235301), and Children’s Hospital Los Angeles (CHLA-19-00348). approved the study protocols and analysis.
Exposures
We obtained data on maternal diet soda and NutraSweet or Equal (aspartame packets) intake from a semiquantitative FFQ completed in the first (11.9 weeks of gestation; SD 3.3) and second (29.2 weeks of gestation; SD 2.4) trimesters. During the study period (1999–2002) these were the most common types of NNS on the market7, and therefore the only sweeteners included in the FFQ. We considered one serving of diet soda to be 12 oz (i.e. one can) and one serving of NutraSweet or Equal to be one packet. We also used the FFQ to calculate sugar sweetened beverages (SSBs), defined as regular (sugar sweetened) soda + fruit drinks. Participants endorsed frequency categories from “never/less than 1 per month” to a maximum of “4 or more cans per day” for diet soda/SSB and from “Never/less than 1 per month” to “4 or more per day” for NutraSweet or Equal packets. Although there was an additional question that asked about consumption of “plain yogurt or yogurt with NutraSweet”, we did not consider this question for our analysis, as plain yogurt typically is unsweetened, and therefore not relevant for quantifying NNS intake. The FFQ was slightly modified from the extensively validated Willett FFQ used in the Nurses’ Health Study and other large cohort studies21–23. We previously calibrated the FFQ during pregnancy by comparing dietary intake values obtained from the FFQ with blood concentrations of several nutrients24. The exact FFQ used can be found on the Project Viva website25. The first trimester FFQ asked about consumption “during this pregnancy” i.e. from the woman’s last menstrual period until the completion of the FFQ. The second trimester FFQ referred to “the previous 3 months”. When diet soda and diet soda + NutraSweet/Equal packets were analyzed separately, results were similar, but effects were slightly stronger using the combined diet soda + NutraSweet/Equal packets exposure. For this reason, we considered the exposure to be diet soda + NutraSweet/Equal packets, which we have termed “NNS” in the remainder of the manuscript. However, we include results using diet soda only in the supplementary data. We also include results of first and second trimester exposures separately in the supplementary data.
Outcome measures
During in-person research visits in infancy (median 6.3 months), early childhood (3.2 years), mid childhood (7.7 years) and early adolescence (12.9 years), trained research staff measured length/height to the nearest 0.1cm and weight to the nearest 0.1kg using a calibrated stadiometer (Shorr Productions, Olney, MD) and weighing scale (SECA 881 mother/baby scale, Hanover, MD for infancy and early childhood visit; Tanita model TBF-300A, Arlington Heights, IL for all other visits). We obtained birth weight and additional data on childhood weight and length/height from medical records, where pediatricians recorded length/height and weight data at each well-child visit during infancy and childhood. As described previously, clinicians used the paper-and-pencil technique for measuring recumbent length for infants <2 years at pediatric clinics. We applied a correction algorithm to account for the systematic overestimation of clinical lengths resulting from this technique26. We calculated BMI as weight in kilograms divided by length or height in meters squared using both research and clinical measures. The mean (SD, min-max) number of BMI measurements per child was 17 (9, 3–58). We calculated age and sex specific BMI z-scores using reference data from the World Health Organization27.
We measured subscapular and triceps skinfold thickness to the nearest 0.1mm using calipers (Holtain Ltd, Crosswell, Wales) and calculated the sum of these two measurements (sum of skinfolds in mm) at the early childhood, mid-childhood, and early adolescence visits. We also measured total body fat mass using whole-body dual energy x-ray absorptiometry (DXA) at the mid-childhood and early adolescence visits using a Hologic model Discovery A (Hologic model Discovery A, Hologic, Bedford, MA) and calculated fat mass index (FMI) as fat mass in kilograms divided by height in meters squared. Quality control was checked daily by scanning a standard synthetic spine to check for machine drift. We used Hologic software QDR version 12.6 for scan analysis. A trained investigator from the Project Viva team checked all scans for movement, positioning, and artifacts, and defined body regions for analysis. Intrarater reliability was high at r=0.99.
Covariates
Mothers reported their age, pre-pregnancy weight, height, race/ethnicity, highest educational attainment and smoking history via questionnaires and interviews at recruitment. Mothers also reported paternal educational attainment and height and weight at enrollment. Participants reported the average weekly hours they spent in physical activity using questions derived from the previously published and validated Physical Activity Scale for the Elderly (PASE)28. Physical activity was classified as walking (“for fun or exercise, including to or from work, but not at work”), light/moderate physical activity (“such as yoga, bowling, stretching classes, and skating, not including walking”), and vigorous recreational activities (“such as jogging, swimming, cycling, aerobics class, skiing, or other similar activities”). We summed walking, light/moderate, and vigorous physical activity for total physical activity. We calculated maternal pre-pregnancy and paternal BMI as weight in kilograms divided by height in meters squared. We categorized maternal smoking history as never, former, smoked during pregnancy, and maternal/paternal educational attainment as college graduate yes or no. We extracted data on parity (nulliparous or multiparous) from outpatient medical records.
Data analysis
We used multivariable linear regression models to examine associations of maternal NNS intake with offspring BMI z-score at each in-person research visit (birth, infancy (median 6.3 months), early childhood (3.2 years), mid-childhood (7.7 years), and early adolescence (12.9 years)) and with offspring body fat in early childhood (sum of skinfolds only), mid-childhood, and early adolescence (sum of skinfolds and FMI). We also examined associations with offspring BMI z-score trajectory by fitting linear mixed-effect models as a function of child age at outcome measurement. This method accounts for the correlation between repeated BMI z-score measures on the same individual at different ages. We included interaction terms between maternal NNS intake and child age to estimate the change in BMI z-score over time associated with higher maternal NNS intake. We used maximum likelihood method of estimation and chose an unstructured working covariance matrix for the random effects variables (intercept and linear slope of child age). In all models, we expressed maternal NNS intake as categorical (quartiles) variables, and adjusted for maternal pre-pregnancy BMI, age, race/ethnicity, educational attainment, parity, pre-pregnancy physical activity, pregnancy smoking history and paternal educational attainment and BMI. We did not adjust for maternal blood glucose or fetal growth because they were likely to fall on the causal pathway29 and therefore attenuate total associations. However, we did conduct a sensitivity analysis whereby we removed GDM cases, which is presented as supplementary data.
Results
Mean ± SD maternal age at enrollment was 32.2 ± 5.0 years with a pre-pregnancy BMI of 24.6 ± 5.2 kg/m2. 70% of participants were white, 11% smoked during pregnancy, 69% were college graduates, and 5% had gestational diabetes (Table 1). Infants were 49% female, with an average gestational age at birth of 39.5 ± 1.9 weeks. Mean NNS intake (diet soda + NutraSweet/Equal packets) averaged across the first and second trimesters was 0.23 ± 0.55 servings/day, where the lowest quartile (Q1) consumed 0.00 ± 0.00 and the highest quartile (Q4) consumed 0.98 ± 0.91 servings/day (Table 1). Mothers who had the highest quartile of NNS intake vs. the lowest quartile (Q4 vs. Q1) had higher pre-pregnancy BMI and were more likely to be of White ethnicity and smoke during pregnancy (Table 1).
Table 1.
Characteristics among 1683 mother-child pairs participating in Project Viva, overall and according to quartile of NNS intake in the first and second trimesters (averaged). Data presented as n (%) or mean (SD).
| Overall | NNS*, average first-second trimester, quartiles | ||||
|---|---|---|---|---|---|
|
| |||||
| Q1 | Q2 | Q3 | Q4 | ||
|
| |||||
| n =1683 | n = 825 | n = 208 | n = 317 | n = 333 | |
| Mother | |||||
| Age, years | 32.2 (5.0) | 31.9 (5.5) | 32.0 (4.4) | 32.4 (4.8) | 32.6 (4.3) |
| Pre-pregnancy BMI, kg/m2 | 24.6 (5.2) | 24.1 (5.0) | 25.1 (5.9) | 24.9 (4.9) | 25.5 (5.5) |
| Pre-pregnancy physical activity, h/w | 9.7 (7.7) | 9.4 (8.1) | 10.3 (7.7) | 9.6 (6.8) | 10.1 (7.6) |
| Race/ethnicity (%) | |||||
| . Black | 233 (14) | 140 (17) | 24 (12) | 45 (14) | 24 (7) |
| . Hispanic | 110 (7) | 55 (7) | 17 (8) | 12 (4) | 26 (8) |
| . Asian | 89 (5) | 46 (6) | 13 (6) | 17 (5) | 13 (4) |
| . White | 1180 (70) | 532 (65) | 148 (71) | 234 (74) | 266 (80) |
| . Other | 66 (4) | 47 (6) | 6 (3) | 9 (3) | 4 (1) |
| College graduate (%) | |||||
| . No | 522 (31) | 274 (33) | 62 (30) | 92 (29) | 94 (28) |
| . Yes | 1156 (69) | 546 (67) | 146 (70) | 225 (71) | 239 (72) |
| Nulliparous (%) | |||||
| . No | 860 (51) | 440 (53) | 92 (44) | 161 (51) | 167 (50) |
| . Yes | 823 (49) | 385 (47) | 116 (56) | 156 (49) | 166 (50) |
| Smoking status (%) | |||||
| . Never | 1144 (68) | 592 (72) | 134 (65) | 214 (68) | 204 (62) |
| . Former | 345 (21) | 156 (19) | 44 (21) | 71 (22) | 74 (22) |
| . Smoked during pregnancy | 187 (11) | 74 (9) | 29 (14) | 32 (10) | 52 (16) |
| Gestational diabetes (%) | 82 (5) | 30 (4) | 10 (5) | 14 (4) | 28 (8) |
| First trimester FFQ | |||||
| Weeks’ gestation | 11.9 (3.2) | 11.9 (3.2) | 11.9 (3.3) | 12.0 (3.4) | 11.6 (3.0) |
| NNS, serv/d | 0.27 (0.62) | 0.00 (0.00) | 0.05 (0.03) | 0.20 (0.14) | 1.11 (0.98) |
| Diet soda, serv/d | 0.23 (0.53) | 0.00 (0.00) | 0.05 (0.03) | 0.20 (0.14) | 0.95 (0.84) |
| Nutrasweet, packets/d | 0.03 (0.25) | 0.00 (0.00) | 0.00 (0.01) | 0.01 (0.03) | 0.16 (0.54) |
| SSB, serv/d | 0.64 (0.86) |
7.5 (1.4) | 7.5 (1.4) | 7.6 (1.2) | 7.5 (1.3) |
| Second trimester FFQ | |||||
| Weeks’ gestation | 29.2 (2.4) | 29.2 (2.5) | 29.3 (2.5) | 29.2 (2.2) | 29.3 (2.3) |
| NNS, serv/d | 0.18 (0.60) | 0.00 (0.00) | 0.02 (0.03) | 0.10 (0.11) | 0.81 (1.13) |
| Diet soda, serv/d | 0.14 (0.50) | 0.00 (0.00) | 0.02 (0.03) | 0.09 (0.11) | 0.62 (0.97) |
| Nutrasweet, packets/d | 0.04 (0.27) | 0.00 (0.00) | 0.00 (0.01) | 0.01 (0.04) | 0.20 (0.58) |
| SSB, serv/d | 0.62 (0.84) | 0.63 (0.84) | 0.63 (0.87) | 0.56 (0.76) | 0.64 (0.89) |
| Average first and second | |||||
| NNS, serv/d | 0.23 (0.55) | 0.00 (0.00) | 0.03 (0.01) | 0.15 (0.06) | 0.98 (0.91) |
| Diet soda, serv/d | 0.19 (0.46) | 0.00 (0.00) | 0.03 (0.01) | 0.14 (0.06) | 0.79 (0.78) |
| Nutrasweet, packets/d | 0.04 (0.24) | 0.00 (0.00) | 0.00 (0.01) | 0.01 (0.03) | 0.19 (0.52) |
| SSB, serv/d | 0.65 (0.79) | 0.69 (0.84) | 0.60 (0.76) | 0.55 (0.64) | 0.64 (0.77) |
| Father | |||||
| Father’s BMI, kg/m2 | 26.4 (4.0) | 26.0 (3.9) | 26.2 (3.8) | 26.4 (3.7) | 27.2 (4.4) |
| Father college graduate, % | |||||
| . No | 497 (32) | 252 (34) | 58 (30) | 94 (31) | 93 (30) |
| . Yes | 1048 (68) | 494 (66) | 133 (70) | 205 (69) | 216 (70) |
| Child | |||||
| Male (%) | 854 (51) | 420 (51) | 94 (45) | 155 (49) | 185 (56) |
| Gestational age at birth, weeks | 39.5 (1.9) | 39.4 (2.0) | 39.6 (2.1) | 39.7 (1.6) | 39.6 (1.7) |
| BMI-z birth | 0.55 (0.96) | 0.52 (0.99) | 0.56 (0.92) | 0.54 (1.01) | 0.63 (0.87) |
| BMI-z 6-month visit | 0.65 (1.02) | 0.60 (1.01) | 0.60 (1.01) | 0.65 (1.02) | 0.80 (1.05) |
| BMI-z early childhood visit | 0.72 (1.00) | 0.66 (0.95) | 0.76 (1.03) | 0.64 (0.94) | 0.94 (1.11) |
| BMI-z mid-childhood visit | 0.55 (1.15) | 0.42 (1.12) | 0.64 (1.14) | 0.57 (1.15) | 0.80 (1.20) |
| BMI-z early adolescence visit | 0.49 (1.23) | 0.42 (1.22) | 0.53 (1.22) | 0.43 (1.26) | 0.71 (1.20) |
| Sum of skinfolds early childhood, mm | 16.7 (4.2) | 16.3 (4.1) | 16.8 (4.2) | 16.5 (3.8) | 17.9 (4.9) |
| Sum of skinfolds mid-childhood, mm | 19.6 (9.3) | 18.5 (8.8) | 20.0 (8.7) | 19.7 (8.9) | 21.9 (10.8) |
| Sum of skinfolds early adolescence, mm | 27.9 (13.3) | 26.8 (13.0) | 28.7 (13.7) | 27.6 (12.6) | 30.3 (14.2) |
| DXA FMI mid-childhood, kg/m2 | 4.4 (1.8) | 4.2 (1.7) | 4.6 (1.9) | 4.4 (1.8) | 4.8 (2.0) |
| DXA FMI early adolescence, kg/m2 | 6.2 (2.9) | 6.1 (2.9) | 6.4 (3.0) | 6.2 (3.0) | 6.6 (3.0) |
SSB: Sugary soda + fruit drinks
NNS: Diet soda + NutraSweet/Equal packets
DXA: Dual-energy x-ray absorptiometry
FMI: Fat mass index
In unadjusted and adjusted multivariable regression models (Table 2, Model 3), NNS intake in the highest vs. lowest quartiles (Q4 vs. Q1) was associated with higher BMI z-score at infancy (β 0.20 units; 95% CI: 0.03, 0.38), early childhood (0.24 units: 0.08, 0.39), mid-childhood (0.31 units; 0.12, 0.50) and early adolescence (0.19 units; −0.02, 0.40), but not birth (0.05 units: −0.12, 0.22). When models were further adjusted for paternal BMI and paternal education, the relationship at the early adolescence timepoint was attenuated (0.14 units; −0.07, 0.35) (Table 2, Model 4).
Table 2.
Associations of maternal NNS intake (quartiles of intake averaged across first and second trimesters) with BMI z-scores at each in-person research visit birth to early adolescence (linear regression)
| NNS* | Model 1 | Model 2 | Model 3 | Model 4 | |
|---|---|---|---|---|---|
| quartile | β (95% CI) | ||||
|
| |||||
| BMI-z birth | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| Q2 | 0.04 (−0.16, 0.23) | 0.07 (−0.12, 0.27) | 0.04 (−0.16, 0.23) | 0.00 (−0.21, 0.21) | |
| Q3 | 0.02 (−0.15, 0.18) | 0.04 (−0.13, 0.21) | 0.01 (−0.16, 0.18) | 0.04 (−0.14, 0.22) | |
| Q4 | 0.11 (−0.06, 0.27) | 0.09 (−0.08, 0.26) | 0.05 (−0.12, 0.22) | 0.03 (−0.15, 0.21) | |
| BMI-z 6 m | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| Q2 | 0.00 (−0.19, 0.19) | 0.00 (−0.20, 0.20) | −0.02 (−0.22, 0.18) | −0.03 (−0.23, 0.17) | |
| Q3 | 0.05 (−0.11, 0.21) | 0.00 (−0.17, 0.17) | −0.02 (−0.19, 0.15) | 0.02 (−0.16, 0.19) | |
| Q4 | 0.20 (0.04, 0.37) | 0.22 (0.05, 0.40) | 0.20 (0.03, 0.38) | 0.20 (0.02, 0.38) | |
| BMI-z early childhood | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| Q2 | 0.10 (−0.08, 0.27) | 0.10 (−0.08, 0.27) | 0.05 (−0.12, 0.23) | 0.03 (−0.15, 0.20) | |
| Q3 | −0.02 (−0.17, 0.13) | −0.03 (−0.19, 0.12) | −0.07 (−0.23, 0.08) | −0.07 (−0.23, 0.08) | |
| Q4 | 0.28 (0.13, 0.43) | 0.31 (0.15, 0.47) | 0.24 (0.08, 0.39) | 0.21 (0.05, 0.37) | |
| BMI-z mid-childhood | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| Q2 | 0.22 (0.00, 0.45) | 0.23 (0.01, 0.45) | 0.16 (−0.06, 0.37) | 0.12 (−0.10, 0.34) | |
| Q3 | 0.15 (−0.04, 0.33) | 0.18 (0.00, 0.37) | 0.11 (−0.08, 0.29) | 0.11 (−0.07, 0.30) | |
| Q4 | 0.38 (0.19, 0.56) | 0.41 (0.22, 0.60) | 0.31 (0.12, 0.50) | 0.21 (0.02, 0.40) | |
| BMI-z early adolescence | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| Q2 | 0.11 (−0.13, 0.35) | 0.06 (−0.19, 0.30) | −0.03 (−0.26, 0.20) | −0.05 (−0.29, 0.18) | |
| Q3 | 0.01 (−0.20, 0.22) | 0.07 (−0.14, 0.28) | −0.05 (−0.25, 0.15) | −0.03 (−0.24, 0.17) | |
| Q4 | 0.29 (0.08, 0.50) | 0.32 (0.10, 0.54) | 0.19 (−0.02, 0.40) | 0.14 (−0.07, 0.35) | |
NNS: Diet soda + NutraSweet/Equal packets
Model 1. Unadjusted
Model 2. Adjusted for maternal age, race/ethnicity, education, parity, pre-pregnancy physical activity, and pregnancy smoking status
Model 3. Model 2 + pre-pregnancy BMI
Model 4. Model 3 + father’s BMI and education
High maternal NNS intake (Q4) was also associated with higher sum of skinfolds in early childhood (β 1.29 mm; 0.60, 1.98), mid-childhood (3.13 mm; 1.59, 4.68), and early adolescence (2.77 mm; 0.45, 5.08) compared with low maternal NNS intake (Q1) (Table 3, Model 3). When models were further adjusted for paternal BMI and paternal education, the relationship at the early adolescence timepoint was slightly attenuated (2.27 mm; −0.06, 4.60) (Table 3, Model 4). High maternal NNS intake was also associated with higher DXA FMI in mid-childhood (β 0.46 kg/m2; 0.12, 0.80) (Table 3, Model 3). However, this association was attenuated after adjusting for paternal factors (0.26 kg/m2; −0.07, 0.59) (Table 3, Model 4).
Table 3.
Associations of maternal NNS intake (quartiles of intake averaged across first and second trimesters) with offspring sum of skinfolds (early childhood, mid-childhood, early adolescence) or dual-energy x-ray absorptiometry fat mass index (mid-childhood and early adolescence) (linear regression)
| NNS* | Model 1 | Model 2 | Model 3 | Model 4 | |
|---|---|---|---|---|---|
| Quartile | β (95% CI) | ||||
|
| |||||
| Sum of skinfolds | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| early childhood, mm | Q2 | 0.49 (−0.25, 1.23) | 0.32 (−0.46, 1.09) | 0.28 (−0.49, 1.06) | 0.21 (−0.59, 1.01) |
| Q3 | 0.24 (−0.41, 0.90) | −0.09 (−0.78, 0.60) | −0.13 (−0.83, 0.56) | −0.03 (−0.73, 0.67) | |
| Q4 | 1.61 (0.97, 2.25) | 1.36 (0.67, 2.05) | 1.29 (0.60, 1.98) | 1.17 (0.47, 1.88) | |
| Sum of skinfolds | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| mid-childhood, mm | Q2 | 1.46 (−0.35, 3.27) | 1.27 (−0.52, 3.05) | 0.84 (−0.92, 2.61) | 0.47 (−1.30, 2.23) |
| Q3 | 1.14 (−0.36, 2.65) | 1.34 (−0.18, 2.86) | 0.91 (−0.60, 2.42) | 0.74 (−0.75, 2.24) | |
| Q4 | 3.41 (1.89, 4.93) | 3.75 (2.19, 5.30) | 3.13 (1.59, 4.68) | 2.33 (0.80, 3.87) | |
| Sum of skinfolds | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| early adolescence, mm | Q2 | 1.90 (−0.68, 4.49) | 1.24 (−1.45, 3.93) | 0.41 (−2.16, 2.98) | 0.10 (−2.50, 2.69) |
| Q3 | 0.74 (−1.48, 2.97) | 1.49 (−0.84, 3.82) | 0.29 (−1.95, 2.53) | 0.37 (−1.89, 2.62) | |
| Q4 | 3.50 (1.22, 5.79) | 4.07 (1.66, 6.49) | 2.77 (0.45, 5.08) | 2.27 (−0.06, 4.60) | |
| DXA FMI | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| mid-childhood, kg/m2 | Q2 | 0.43 (0.01, 0.84) | 0.32 (−0.09, 0.73) | 0.24 (−0.16, 0.64) | 0.17 (−0.21, 0.56) |
| Q3 | 0.21 (−0.12, 0.54) | 0.24 (−0.09, 0.56) | 0.16 (−0.17, 0.48) | 0.13 (−0.19, 0.45) | |
| Q4 | 0.61 (0.27, 0.94) | 0.59 (0.24, 0.93) | 0.46 (0.12, 0.80) | 0.26 (−0.07, 0.59) | |
| DXA FMI | Q1 | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) | 0.0 (ref) |
| early adolescence, kg/m2 | Q2 | 0.36 (−0.34, 1.05) | 0.17 (−0.53, 0.87) | −0.03 (−0.70, 0.63) | −0.09 (−0.76, 0.58) |
| Q3 | 0.08 (−0.51, 0.67) | 0.13 (−0.47, 0.74) | −0.16 (−0.74, 0.41) | −0.09 (−0.67, 0.48) | |
| Q4 | 0.55 (−0.04, 1.14) | 0.57 (−0.04, 1.18) | 0.28 (−0.30, 0.86) | 0.07 (−0.52, 0.66) | |
NNS: Diet soda + NutraSweet/Equal packets
Model 1. Unadjusted
Model 2. Adjusted for maternal age, race/ethnicity, education, parity, pre-pregnancy physical activity, and pregnancy smoking status
Model 3. Model 2 + pre-pregnancy BMI
Model 4. Model 3 + father’s BMI and education
DXA: Dual-energy x-ray absorptiometry
FMI: Fat mass index
In adjusted mixed-effect models, there was a positive interaction between the maternal NNS intake-offspring BMI z-score relationship and child age (0.03 units/year; 0.01–0.05; Pinteraction with age: <0.0001) (Table 4, Model 4). This implies that the strength of the association between maternal NNS intake and offspring BMI z-score increased as the children aged. For example, the estimates of associations between highest quartile vs. lowest quartile of maternal NNS intake and offspring BMI z-score were 0.01 units (−0.11, 0.13) at birth, 0.10 units (−0.01, 0.22) at 3 years, 0.23 units (0.10, 0.37) at 7 years, 0.39 units (0.19, 0.59) at 12 years and 0.45 units (0.23, 0.68) at 14 years, respectively (Figure 1).
Table 4.
Associations of NNS* (quartiles of intake averaged across first and second trimesters) quartiles with BMI trajectory from birth to 18 years#
| NNS intake | |||||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | ||
| β (95% CI) | |||||
|
| |||||
| Model 1 | Main effect | Ref | −0.24 (−1.83, 1.35) | 0.23 (−1.12, 1.59) | 1.00 (−0.34, 2.34) |
| Interaction | 0.01 (−0.004, 0.03) | 0.01 (−0.003, 0.03) | 0.03 (0.01, 0.04) | ||
| Model 2 | Main effect | Ref | 0.03 (−1.60, 1.67) | 0.13 (−1.28, 1.54) | 1.22 (−0.19, 2.63) |
| Interaction | 0.01 (−0.007, 0.03) | 0.02 (−0.001, 0.03) | 0.03 (0.01, 0.04) | ||
| Model 3 | Main effect | Ref | −0.35 (−1.99, 1.30) | −0.33 (−1.75, 1.09) | 0.56 (−0.86, 1.99) |
| Interaction | 0.01 (−0.006, 0.03) | 0.02 (−0.001, 0.03) | 0.03 (0.01, 0.05) | ||
| Model 4 | Main effect | Ref | −0.29 (−2.00, 1.42) | −0.11 (−1.57, 1.35) | 0.10 (−1.37, 1.57) |
| Interaction | 0.01 (−0.01, 0.03) | 0.02 (−0.002, 0.03) | 0.03 (0.01, 0.05) | ||
NNS: Diet soda + NutraSweet/Equal packets
Expressed as BMI z-score SD units/year of age
Model 1. Unadjusted
Model 2. Adjusted for maternal age, race/ethnicity, education, parity, pre-pregnancy physical activity, and pregnancy smoking status
Model 3. Model 2 + pre-pregnancy BMI
Model 4. Model 3 + father’s BMI and education
Figure 1. Predicted linear mixed effects models of BMI z-score and NNS intake (quartiles of intake averaged across first and second trimesters) as a function of age (birth to 18 years), adjusted for maternal age, race/ethnicity, education, parity, pre-pregnancy physical activity, pregnancy smoking status, and pre-pregnancy BMI, and father’s BMI and education.
Between months 48–216, P<0.05 for Q4 vs. Q1. There were no differences between other quartiles.
*NNS: Diet soda + NutraSweet/Equal packets
Adjusted for maternal age, race/ethnicity, education, parity, pre-pregnancy physical activity, pregnancy smoking status, and pre-pregnancy BMI, and father’s BMI and education
When split by trimester and considering diet soda alone (i.e. no NutraSweet/Equal) results were similar, although first trimester exposures had stronger associations with offspring BMI z-score (Supplementary Table 1) and body fat (Supplementary Table 2) than second trimester exposures. Removing gestational diabetes cases from the analysis did not materially change the results (Supplementary Tables 3 and 4).
Discussion
This study adds to a growing body of evidence that suggests that maternal NNS intake during pregnancy increases childhood BMI. However, unlike previous studies that examined these associations at one timepoint only, we examined childhood BMI z-score longitudinally, from birth to age 18 years, as well as two different measures of body fat (sum of skinfolds and FMI). We found that maternal NNS intake, averaged across the first and second trimesters, was associated with increased offspring BMI z-score and body fat from infancy to 18 years, and that the effect became more pronounced with age.
Three previous studies have examined the relationship between maternal NNS consumption and childhood BMI in humans. Azad et al. (2016) found that, compared with no intake, daily maternal diet soda intake in the second or third trimester was associated with a 0.20 unit increase in infant BMI z-score at 1 year of age in a Canadian cohort of 2686 mother-infant pairs (the CHILD Study)8. Zhu et al. (2017) found that daily second trimester artificially sweetened beverage intake was associated with a 0.59 unit increase in BMI z-score at 7 years amongst 918 mother-child pairs in the Danish National Birth Cohort16. In contrast, in a prior analysis of 1078 mother-child pairs from the Project Viva cohort, Gillman et al. (2017) did not find a relationship between maternal diet soda consumption in pregnancy and childhood BMI z-score or body fat at mid-childhood5. While the primary aim of that analysis was to assess the effects of maternal SSB consumption on mid-childhood BMI and adiposity, the authors also examined the effects of maternal diet soda intake as a comparison to the effects of maternal SSB intake. The reasons for the discrepancy in results between the current analysis and those reported by Gillman et al. are differences in the definition of the exposure and the timepoint of the outcome measure. Gillman et al. considered only the effect of diet soda (not including NutraSweet/Equal packets) consumption in the second trimester (not first trimester or averaged first/second trimester), and looked only at BMI z-score and FMI at the mid-childhood (7.7 year) timepoint. Our results similarly show no association of second trimester maternal diet soda intake with BMI z-score or FMI at the mid-childhood timepoint. However, when diet soda was combined with NutraSweet/Equal, and first and second trimester exposures were averaged, there was indeed an association between NNS exposure and increased BMI z-score and FMI in mid-childhood. Diet soda alone was also associated with higher BMI z-score in mid-childhood, as long as only the first trimester or the averaged first-second trimesters were considered. This suggests that, in our cohort, first trimester NNS exposure was more important than second trimester exposure. Taken together, all prospective cohort studies that have published on the effects of maternal NNS intake during pregnancy and childhood BMI (CHILD Study, Danish National Birth Cohort Study, and now Project Viva) have demonstrated a positive relationship between maternal NNS intake during pregnancy and higher childhood BMI.
Given the narrow range of NNS consumption within our cohort, we opted to compare the highest quartile of NNS intake (Q4) with the lowest quartile of NNS intake (Q1) in our analyses. Since Q4 consumed 0.98 ± 0.88 servings/day while Q1 consumed 0.00 ± 0.00 servings/day, it is reasonable to summarize these results as the difference between approximately daily consumption (0.98 ± 0.88 servings/day in Q4) vs. no consumption (0 servings/day in Q1). Therefore, our results indicate that approximate daily NNS consumption during pregnancy was associated with a 0.20 unit higher BMI z-score in infancy (median 6.3 months), 0.21 unit higher BMI z-score in early childhood (3.2 years), and 0.21 unit higher BMI z-score in mid-childhood (7.7 years), compared with no NNS intake during pregnancy. These effects are slightly smaller than those reported by Azad et al. 2016 (0.20 unit increase in BMI z-score at 1 year) and Zhu et al. 2017 (0.59 unit increase in BMI z-score at 7 years), which could be due to population differences or differences in the confounders adjusted for across studies. For example, neither of these prior studies adjusted for paternal BMI or educational attainment. Our BMI trajectory analysis also demonstrated differences in childhood BMI trajectory from birth to 18 years between the highest and lowest quartiles of maternal NNS intake. This relationship strengthened with age, and by 18 years, the difference in BMI z-score between Q1 and Q4 was 0.58 units (p = 0.0001) – whereby Q1 had a predicted BMI z-score of 0.60 and Q4 had a predicted BMI z-score of 1.18 (depicted in Figure 1). Since childhood overweight is defined as a BMI z-score >130, this result indicates that no maternal NNS intake was associated with having a normal weight child at 18 years, while approximately 1 serving/day maternal NNS intake was associated with having an overweight child at 18 years. Given that 20% (333/1683) of our cohort and up to 30% of pregnant women in other cohorts8,16 report daily NNS consumption, these results are meaningful for population health.
We also found that, compared with no NNS intake (Q1), daily NNS intake (Q4) was associated with higher sum of skinfolds in early childhood (1.17 mm), mid-childhood (2.33 mm), and early adolescence (2.27 mm). Similar results were found with FMI, whereby daily NNS intake was associated with a 0.26 kg/m2 higher FMI in mid-childhood compared with no NNS intake. Several animal and cell culture studies have similarly reported effects of maternal NNS intake on offspring adiposity, including a recent rat study where low-dose maternal aspartame and stevia consumption increased offspring weight gain and adiposity from birth to 18 weeks of age31. Treatment of mouse and human precursor cells with the NNSs saccharin and acesulfame-K also enhances adipogenesis in culture32. Taken together, these results suggest that maternal NNS intake not only increases BMI z-score throughout childhood but also adiposity, which is a stronger predictor of metabolic disease than BMI alone19.
There are several potential mechanisms that could explain our findings. Prenatal programming may be mediated by epigenetic changes33. It is also possible that mothers continued to consume NNS during lactation and that these NNS were transmitted via breast milk, as has been demonstrated previously7. Differences in BMI/adiposity could also reflect the child’s own consumption of NNS or sugar, which may be proportional to the mother’s NNS consumption. Rodent studies also demonstrated that prenatal exposure to NNSs increase offspring sweet taste preference34. It is therefore possible that children exposed to NNS in utero are more likely to eat sweet, calorie-dense foods, increasing their risk for obesity. Furthermore, although we controlled for several potential confounders such as maternal and paternal BMI, race/ethnicity, and education level in the analysis, there is still the possibility of residual confounding35.
Strengths of this study include the relatively large sample size, classification of exposure in both the first and second trimesters, the combination of beverage and packaged sources of NNS, the wealth of detailed information collected (including covariates), the inclusion of both BMI and fat mass, and the prospective follow-up of BMI z-score from birth to 18 years using a combination of clinical and research measures. Limitations are that most of the participants were college educated and of white ethnicity, limiting the generalizability of the results. We also didn’t consider post-exposure factors such as child dietary habits or physical activity, which could potentially modify relationships between maternal NNS consumption and childhood growth and adiposity. Further, we did not distinguish between types of NNS. Diet soda in 1999–2002 consisted of a mix of various NNS, but most commonly aspartame and acesulfame-potassium. NutraSweet/Equal are also aspartame7. Since the early 2000s, other NNS such as sucralose and stevia have been popularized. Given their various chemical structures and pharmacokinetic properties36, different NNS are likely to exert different effects on offspring. Future studies should examine other types of NNS, as well as possible mechanisms.
In conclusion, we have demonstrated that prenatal exposure to NNSs is associated with higher BMI z-score and body fat longitudinally throughout childhood, from birth to age 18, an effect that becomes more pronounced with age. While these results are observational, and therefore conclusions about causality cannot be drawn, our study adds to a growing body of evidence that maternal NNS intake during pregnancy is associated with increased risk of childhood obesity.
Supplementary Material
Acknowledgements:
We thank the participants and staff of Project Viva.
This work was supported by the US National Institutes of Health (R01 HD034568, UH3 OD023286).
Footnotes
Competing interests: None declared.
REFERENCES
- 1.Centers for Disease Control and Prevention. Obesity and Overweight Statistics in the United States 2015–2016. 2019.https://www.cdc.gov/nchs/fastats/obesity-overweight.htm (accessed 21 Nov2019).
- 2.Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of Childhood and Adult Obesity in the United States, 2011–2012. JAMA 2014; 311: 806–814. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ward ZJ, Long MW, Resch SC, Giles CM, Cradock AL, Gortmaker SL. Simulation of Growth Trajectories of Childhood Obesity into Adulthood. N Engl J Med 2017; 377: 2145–2153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Buscot M-J, Thomson RJ, Juonala M, Sabin MA, Burgner DP, Lehtimäki T et al. Distinct child-to-adult body mass index trajectories are associated with different levels of adult cardiometabolic risk. Eur Heart J 2018; 39: 2263–2270. [DOI] [PubMed] [Google Scholar]
- 5.Gillman MW, Rifas-Shiman SL, Fernandez-Barres S, Kleinman K, Taveras EM, Oken E. Beverage Intake During Pregnancy and Childhood Adiposity. Pediatrics 2017; : e20170031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Symonds ME, Sebert SP, Hyatt MA, Budge H. Nutritional programming of the metabolic syndrome. Nat Rev Endocrinol 2009; 5: 604–610. [DOI] [PubMed] [Google Scholar]
- 7.Sylvetsky AC, Rother KI. Trends in the Consumption of Low-Calorie Sweeteners. Physiol Behav 2016; 164: 446–450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Azad MB, Sharma AK, de Souza RJ, Dolinsky VW, Becker AB, Mandhane PJ et al. Association Between Artificially Sweetened Beverage Consumption During Pregnancy and Infant Body Mass Index. JAMA Pediatr 2016; 170: 662–670. [DOI] [PubMed] [Google Scholar]
- 9.Pepino MY, Bourne C. Non-nutritive sweeteners, energy balance, and glucose homeostasis. Curr Opin Clin Nutr Metab Care 2011; 14: 391–395. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Davidson TL, Martin AA, Clark K, Swithers SE. Intake of high-intensity sweeteners alters the ability of sweet taste to signal caloric consequences: implications for the learned control of energy and body weight regulation. Q J Exp Psychol 2006. 2011; 64: 1430–1441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Blundell JE, Hill AJ. Paradoxical effects of an intense sweetener (aspartame) on appetite. Lancet Lond Engl 1986; 1: 1092–1093. [DOI] [PubMed] [Google Scholar]
- 12.Suez J, Korem T, Zeevi D, Zilberman-Schapira G, Thaiss CA, Maza O et al. Artificial sweeteners induce glucose intolerance by altering the gut microbiota. Nature 2014; 514: 181–186. [DOI] [PubMed] [Google Scholar]
- 13.Plows JF, Ponnampalam AP, Vickers MH, Reynolds CM. Artificial Sweeteners during Pregnancy-Impact on Maternal Metabolic Health. In: Reproductive Sciences. Sage Publications, 2018, pp 99A–100A. [Google Scholar]
- 14.Logan KM, Emsley RJ, Jeffries S, Andrzejewska I, Hyde MJ, Gale C et al. Development of Early Adiposity in Infants of Mothers With Gestational Diabetes Mellitus. Diabetes Care 2016; 39: 1045–1051. [DOI] [PubMed] [Google Scholar]
- 15.Logan KM, Gale C, Hyde MJ, Santhakumaran S, Modi N. Diabetes in pregnancy and infant adiposity: systematic review and meta-analysis. Arch Dis Child Fetal Neonatal Ed 2017; 102: F65–F72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zhu Y, Olsen SF, Mendola P, Halldorsson TI, Rawal S, Hinkle SN et al. Maternal consumption of artificially sweetened beverages during pregnancy, and offspring growth through 7 years of age: a prospective cohort study. Int J Epidemiol 2017; 46: 1499–1508. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Crume TL, Ogden L, Daniels S, Hamman RF, Norris JM, Dabelea D. The Impact of In Utero Exposure to Diabetes on Childhood Body Mass Index Growth Trajectories: The EPOCH Study. J Pediatr 2011; 158: 941–946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Johnson W, Choh AC, Lee M, Towne B, Czerwinski SA, Demerath EW. Characterization of the Infant BMI Peak: Sex Differences, Birth Year Cohort Effects, Association with Concurrent Adiposity, and Heritability. Am J Hum Biol Off J Hum Biol Counc 2013; 25: 378–388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Vanderwall C, Randall Clark R, Eickhoff J, Carrel AL. BMI is a poor predictor of adiposity in young overweight and obese children. BMC Pediatr 2017; 17. doi: 10.1186/s12887-017-0891-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Oken E, Baccarelli AA, Gold DR, Kleinman KP, Litonjua AA, De Meo D et al. Cohort profile: project viva. Int J Epidemiol 2015; 44: 37–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Rifas-Shiman SL, Rish-Edwards JW, Kleinman KP, Oken E, Gillman MW. Dietary Quality during Pregnancy Varies by Maternal Characteristics in Project Viva: A US Cohort. J Am Diet Assoc 2009; 109: 1004–1011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hu FB, Stampfer MJ, Manson JE, Rimm E, Colditz GA, Rosner BA et al. Dietary fat intake and the risk of coronary heart disease in women. N Engl J Med 1997; 337: 1491–1499. [DOI] [PubMed] [Google Scholar]
- 23.Rimm EB, Giovannucci EL, Stampfer MJ, Colditz GA, Litin LB, Willett WC. Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am J Epidemiol 1992; 135: 1114–1126; discussion 1127–1136. [DOI] [PubMed] [Google Scholar]
- 24.Fawzi WW, Rifas-Shiman SL, Rich-Edwards JW, Willett WC, Gillman MW. Calibration of a semi-quantitative food frequency questionnaire in early pregnancy. Ann Epidemiol 2004; 14: 754–762. [DOI] [PubMed] [Google Scholar]
- 25.Project Viva | A Study of Health for The Next Generation. https://www.hms.harvard.edu/viva/data-collection-by-domain.html (accessed 20 Nov2020).
- 26.Rifas-Shiman SL, Rich-Edwards JW, Scanlon KS, Kleinman KP, Gillman MW. Misdiagnosis of overweight and underweight children younger than 2 years of age due to length measurement bias. MedGenMed Medscape Gen Med 2005; 7: 56. [PMC free article] [PubMed] [Google Scholar]
- 27.WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards based on length/height, weight and age. Acta Paediatr Oslo Nor 1992. Suppl 2006; 450: 76–85. [DOI] [PubMed] [Google Scholar]
- 28.Washburn RA, McAuley E, Katula J, Mihalko SL, Boileau RA. The physical activity scale for the elderly (PASE): evidence for validity. J Clin Epidemiol 1999; 52: 643–651. [DOI] [PubMed] [Google Scholar]
- 29.Ananth CV, Schisterman EF. Confounding, Causality and Confusion: The Role of Intermediate Variables in Interpreting Observational Studies in Obstetrics. Am J Obstet Gynecol 2017; 217: 167–175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.de Onis M, Onyango AW, Borghi E, Siyam A, Nishida C, Siekmann J. Development of a WHO growth reference for school-aged children and adolescents. Bull World Health Organ 2007; 85: 660–667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Nettleton JE, Cho NA, Klancic T, Nicolucci AC, Shearer J, Borgland SL et al. Maternal low-dose aspartame and stevia consumption with an obesogenic diet alters metabolism, gut microbiota and mesolimbic reward system in rat dams and their offspring. Gut 2020. doi: 10.1136/gutjnl-2018-317505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Simon BR, Parlee SD, Learman BS, Mori H, Scheller EL, Cawthorn WP et al. Artificial Sweeteners Stimulate Adipogenesis and Suppress Lipolysis Independently of Sweet Taste Receptors. J Biol Chem 2013; 288: 32475–32489. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Gicquel C, El-Osta A, Le Bouc Y. Epigenetic regulation and fetal programming. Best Pract Res Clin Endocrinol Metab 2008; 22: 1–16. [DOI] [PubMed] [Google Scholar]
- 34.Zhang G-H, Chen M-L, Liu S-S, Zhan Y-H, Quan Y, Qin Y-M et al. Effects of mother’s dietary exposure to acesulfame-K in Pregnancy or lactation on the adult offspring’s sweet preference. Chem Senses 2011; 36: 763–770. [DOI] [PubMed] [Google Scholar]
- 35.Liang W, Zhao Y, Lee AH. An Investigation of the Significance of Residual Confounding Effect. BioMed Res Int 2014; 2014. doi: 10.1155/2014/658056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Magnuson BA, Carakostas MC, Moore NH, Poulos SP, Renwick AG. Biological fate of low-calorie sweeteners. Nutr Rev 2016; 74: 670–689. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.