Diabetologia (2019) 62:915–925 https://doi.org/10.1007/s00125-019-4842-0 ARTICLE A reduction in sedentary behaviour in obese women during pregnancy reduces neonatal adiposity: the DALI randomised controlled trial Mireille N. M. van Poppel 1,2 & David Simmons 3,4 & Roland Devlieger 5,6 & F. Andre van Assche 5,6 & Goele Jans 5,6 & Sander Galjaard 5,6,7 & Rosa Corcoy 8,9 & Juan M. Adelantado 8 & Fidelma Dunne 10,11 & Jürgen Harreiter 12 & Alexandra Kautzky-Willer 12 & Peter Damm 13 & Elisabeth R. Mathiesen 13 & Dorte M. Jensen 14,15,16 & Lise-Lotte Andersen 14,15,16 & Mette Tanvig 14,15,16,17 & Annunziata Lapolla 18 & Maria G. Dalfra 18 & Alessandra Bertolotto 19 & Ewa Wender-Ozegowska 20 & Agnieszka Zawiejska 20 & David Hill 21 & Frank J. Snoek 22 & Judith G. M. Jelsma 1 & Gernot Desoye 23 Received: 30 October 2018 / Accepted: 6 February 2019 / Published online: 6 March 2019 # The Author(s) 2019 Abstract Aims/hypothesis Offspring of obese women are at increased risk of features of the metabolic syndrome, including obesity and diabetes. Lifestyle intervention in pregnancy might reduce adverse effects of maternal obesity on neonatal adiposity. Methods In the Vitamin D And Lifestyle Intervention for Gestational Diabetes Mellitus (GDM) Prevention (DALI) lifestyle trial, 436 women with a BMI ≥29 kg/m2 were randomly assigned to counselling on healthy eating (HE), physical activity (PA) or HE&PA, or to usual care (UC). In secondary analyses of the lifestyle trial, intervention effects on neonatal outcomes (head, abdominal, arm and leg circumferences and skinfold thicknesses, estimated fat mass, fat percentage, fat-free mass and cord blood leptin) were assessed using multilevel regression analyses. Mediation of intervention effects by lifestyle and gestational weight gain was assessed. Results Outcomes were available from 334 neonates. A reduction in sum of skinfolds (−1.8 mm; 95% CI −3.5, −0.2; p = 0.03), fat mass (−63 g; 95% CI −124, −2; p = 0.04), fat percentage (−1.2%; 95% CI −2.4%, −0.04%; p = 0.04) and leptin (−3.80 μg/l; 95% CI −7.15, −0.45; p = 0.03) was found in the HE&PA group, and reduced leptin in female neonates in the PA group (−5.79 μg/l; 95% CI −11.43, −0.14; p = 0.05) compared with UC. Reduced sedentary time, but not gestational weight gain, mediated intervention effects on leptin in both the HE&PA and PA groups. Conclusions/interpretation The HE&PA intervention resulted in reduced adiposity in neonates. Reduced sedentary time seemed to drive the intervention effect on cord blood leptin. Implications for future adiposity and diabetes risk of the offspring need to be elucidated. Trial registration ISRCTN70595832. Keywords Lifestyle intervention . Mediation . Neonatal adiposity . Randomised controlled trial Professor F. A. van Assche, who contributed to this research, died on 30 Abbreviations July 2018, before publication of this work. DALI Vitamin D And Lifestyle Intervention for Electronic supplementary material The online version of this article Gestational Diabetes Mellitus (GDM) Prevention (https://doi.org/10.1007/s00125-019-4842-0) contains peer-reviewed but HE Counselling on healthy eating (intervention group) unedited supplementary material, which is available to authorised users. HE&PA Counselling on healthy eating and physical activity (intervention group) * Mireille N. M. van Poppel MI Motivational interviewing Mireille.van-poppel@uni-graz.at PA Counselling on physical activity (intervention group) UC Usual care (intervention group) Extended author information available on the last page of the article 916 Diabetologia (2019) 62:915–925 Introduction In the pan-European Vitamin D And Lifestyle Intervention for Gestational Diabetes Mellitus (GDM) Prevention (DALI) Globally, an estimated 41 million children under 5 years of trial, motivational interviewing (MI)-based counselling on age were either overweight or obese in 2014, which is a major healthy eating (HE), physical activity (PA) and a combination public health concern [1]. of the two (HE&PA) was compared with usual care (UC) Obesity during the lifetime, including childhood, tracks [18–20]. Primary outcomes were gestational weight gain, back to the early postnatal period [2] and, as recently reported, fasting glucose and insulin sensitivity. We previously reported even back to birth: the Hyperglycemia and Adverse improvements in lifestyle behaviour in all three intervention Pregnancy Outcome (HAPO) follow-up study, including groups, in addition to a substantial reduction in gestational 4832 children at a mean age of 11.4 years, demonstrated the weight gain in the HE&PA group, but no changes in fasting association of adiposity at birth with adiposity in late child- glucose or insulin sensitivity [20]. hood [3]. Furthermore, a recent study showed that being large Here we tested whether the lifestyle interventions altered for gestational age at birth is related to an increased risk of neonatal anthropometry and cord blood leptin, as a marker of adolescent obesity [4]. Therefore, reducing adiposity at birth adiposity [21], both secondary outcomes of the DALI lifestyle is a key target for the prevention of obesity later in life. trial. We also investigated whether changes in neonatal adi- Maternal obesity and excessive gestational weight gain in posity were primarily mediated through changes in lifestyle, pregnancy are implicated in the development of neonatal gestational weight gain, or both. The findings would allow obesity [5–8]. Lifestyle intervention in pregnant, obese targeted counselling of pregnant women to improve the women might reduce the adverse effects of maternal obesity long-term weight of their offspring and, hence, address a ma- on offspring adiposity, possibly associated with a reduction jor public health concern. in gestational weight gain. Two previous studies have shown that treatment of women with mild gestational diabetes re- sults in a reduction in neonatal adiposity [9, 10], although it Methods was not maintained into childhood [11, 12]. Disappointingly, lifestyle trials in healthy pregnant obese women did not find Design and participants such effects on neonatal adiposity [13–15], possibly owing to insufficient change in lifestyle behaviours or minimal Originally, the DALI study was designed as a trial with a changes in gestational weight gain [16, 17]. 2 × (2 × 2) factorial design, including a vitamin D trial and a Diabetologia (2019) 62:915–925 917 lifestyle trial [18]. The two trials were not conducted in paral- In the intervention groups, participants were assigned to a lel as originally planned but, because of logistical reasons, the single coach, with whom they discussed five physical activity vitamin D trial started after the lifestyle trial had finished. The and/or seven healthy eating messages, depending on group current paper describes secondary analyses of the DALI life- allocation, and were advised to keep gestational weight gain style trial, a multicentre parallel randomised trial conducted in below 5 kg [18]. Most lifestyle coaches had a background in nine European countries (Austria, Belgium, Denmark behavioural change, healthy eating and/or physical activity. [Odense, Copenhagen], Ireland, Italy [Padua, Pisa], the Prior to the start of the study, a 2 day central training course Netherlands, Poland, Spain and the UK) during 2012–2015. in Cambridge (UK) was offered to the coaches of all sites, The study was prospectively registered as an RCT on 21 which was led by experienced MI trainers. This was repeated November 2011 (ISRCTN70595832). Local ethics committee a few months later to review the coaches’ MI competency, approval and written informed consent of all women was ob- share experiences and receive feedback on a role play. The tained. Pregnant women with a pre-pregnancy BMI ≥29 kg/ DALI coaches received in total 32 h of MI training. Their MI m2, <20 weeks of gestation, a singleton pregnancy and age ≥ skills in the trial were evaluated by external MI experts [23]. 18 years were invited to participate. Exclusions included di- MI-guided coaching of the trial participants took place during agnosis with early gestational diabetes [22], pre-existing dia- five face-to-face sessions of 30–45 min each, alternated with betes and chronic medical conditions. The numbers of women up to four optional telephone calls. In the UC group, partici- excluded and included are shown in the flow chart (Fig. 1). pants received no DALI interventions. Randomisation, masking and interventions Outcomes Women were randomised to HE&PA, HE, PA or UC, using a Neonatal anthropometry Neonatal weight and length were computerised random number generator, pre-stratified for site. measured at birth and, within 48 h, head, abdomen, upper- Staff involved with measurements, but not participants, were and lower-arm and upper- and lower-leg circumferences were blinded to the intervention. measured. Research nurses received central training for the Approached (N = 2009) Declined participation (n = 1233) Excluded (n = 137) Withdrew consent (n = 7) Consented and assessed for inclusion Excluded (n = 196): (n = 639) • GDM (n = 170) • BMI < 29 kg/m2 (n = 8) • Gestation ≥ 20 weeks (n = 8) • No baseline OGTT (n = 4) • Other reasons (n = 6) Randomised (n = 436) Allocated to UC Allocated to HE&PA Allocated to HE Allocated to PA (n = 105) (n = 108) (n = 113) (n = 110) Lost to follow-up (n = 11): Lost to follow-up (n = 15): Lost to follow-up (n = 10): Lost to follow-up (n = 12): • Withdrew consent (n = 1) • Withdrew consent (n = 1) • Too busy/no time (n = 2) • Withdrew consent (n = 2) • Mental illness (n = 1) • Too busy/no time (n = 2) • Too many blood tests (n = 1) • Too busy/no time (n = 2) • Moved away (n = 2) • Family illness (n = 1) • Abortion/miscarriage (n = 1) • Medical reason (n =1) • Unknown (n = 7) • Abortion/miscarriage/fetal death (n = 3) • Moved away (n = 2) • Being active too difficult (n = 1) • Moved away (n = 1) • Unknown (n = 4) • Unknown (n = 6) • Unknown (n = 7) Live born neonates (n = 94) Live born neonates (n = 93) Live born neonates (n =103) Live born neonates (n = 98) Data on anthropometrics or cord Data on anthropometrics or cord Data on anthrop ometrics or cord Data on anthropometrics or cord blood leptin (n = 80) blood leptin (n = 80) blood leptin (n = 92) blood leptin (n = 82) Fig. 1 Eligibility, randomisation and follow-up of women and neonates in the DALI lifestyle trial. GDM, gestational diabetes mellitus 918 Diabetologia (2019) 62:915–925 measurements at the beginning of the trial and participated in a Patient involvement standardisation meeting within 6 months after trial start. Skinfold thickness was measured at four sites, i.e. triceps, Representatives of the target group were interviewed in the de- subscapular, suprailiac and quadriceps, with a Harpenden velopmental stage of the trial about their preferences of interven- skinfold calliper [18], and values summed. Each skinfold tion content, modality, frequency and location [27]. Patients were measurement was repeated once and if a difference of more not involved in the actual conduct of the study, but study partic- than 0.2 mm was registered a third measurement was per- ipants provided feedback on the burden of the intervention and formed and the average of the three was taken. Time between their experiences with the study in general, as part of a process birth and measurements was registered (in h). Estimated fat evaluation [23]. Patient organisations are actively involved in the mass (in g) was calculated using a validated equation for ne- dissemination of the results to the lay public. onates [24], with all neonates classified as ‘non-Hispanic’. Estimated fat-free mass (in g) was calculated as total body Statistical analyses mass minus estimated fat mass. Estimated fat percentage was calculated by dividing estimated fat mass by total body Data presented in this paper are secondary outcomes of the mass × 100. DALI lifestyle trial [18]. All neonatal outcomes presented in this paper were pre-specified in the protocol, with the excep- Cord blood leptin concentration Venous cord blood samples tion of estimated fat mass, fat-free mass and fat percentage, were taken immediately after delivery. Locally, blood samples which were calculated based on pre-specified variables. were stored at −20°C or colder until transported frozen to the Neonatal loss to follow-up was defined as no data on both central trial laboratory in Graz, Austria. There, they were neonatal anthropometry and cord blood leptin. Differences stored at −70°C for 3 years maximum (the first women includ- between the study sample and participants with loss to ed) and 2 months minimum (last women included). Leptin follow-up and differences between the intervention groups concentrations were quantified by solid-phase sandwich and the UC group were tested using Student’s t test for con- ELISA (E05–086-96; EIASON, Graz, Austria), according to tinuous variables or χ2 tests for categorical variables. the manufacturer’s instructions. Analytical sensitivity was To test for intervention effects, multilevel analyses were un- 1.0 ng/ml; intra- and inter-assay coefficients of variation dertaken with a two level structure: individual and site. A mod- (low/high concentrations) were 6.0%/6.9% and 11.6%/8.7%, ified intention-to-treat approach was used, without any imputa- respectively. tion of missing data. A sensitivity analysis was conducted with multiple imputation under missing-at-random assumptions. Maternal height, weight and lifestyle characteristics Maternal Analyses of neonatal anthropometric variables were always ad- height was determined at baseline with a stadiometer (SECA justed for the time after birth (in h). Additional adjustment of 206; SECA, Birmingham, UK). Women were weighed on analyses by parity, pre-pregnancy BMI, education, smoking, calibrated electronic scales (SECA 888 and 877) at baseline sex, gestational age at birth and mode of delivery did not change and at 24–28 weeks and 35–37 weeks of gestation. BMI was the results and are therefore not shown. Effect dependency by sex calculated as weight (in kg) divided by the square of height (in was assessed for all outcomes. An interaction with p < 0.10 was m). Gestational weight gain was defined as the change in considered relevant and, when present, analyses were performed objectively measured weight from baseline to 24–28 weeks for both sexes separately. All analyses were performed with IBM and from baseline to 35–37 weeks. SPSS, version 20 (IBM, Armonk, NY, USA), with a two-sided p At baseline and at 24–28 and 35–37 weeks, maternal phys- value below 0.05 considered significant. No adjustments for mul- ical activity was assessed with the validated Pregnancy tiple comparisons were made. Physical Activity Questionnaire [25], which assesses the time Post hoc power calculations showed that a difference of spent sedentary (watching TV or video, sitting and reading, 2.0 mm (SD 5.0) in sum of skinfolds could be found with an talking or on the phone) and in light, moderate or vigorous α value of 5% and power of 80% with 98 women per group, physical activity measured in metabolic equivalent of task and a difference of 1.5% (SD 3.5) in estimated fat percentage (MET) h/week. Nutrition was assessed using a bespoke, short could be found with 85 women in each group. food-frequency questionnaire covering key foods, linked to Mediation by lifestyle factors was assessed in multiple parallel the intervention messages [26]. mediation models [28, 29]. This allows the determination of the lifestyle factor with the strongest indirect effect [28]. The model Covariates Information on possible covariates was collected in was limited to the six lifestyle factors that changed significantly the baseline questionnaire or from medical files: parity (nul- due to the interventions (moderate-to-vigorous physical activity, liparous vs multiparous), education (low vs high), smoking sedentary behaviour, sugared drink consumption, vegetable con- status (yes vs no), pre-pregnancy BMI, mode of delivery sumption, carbohydrate intake and portion size) [20]. (Caesarean vs vaginal delivery) and gestational age at birth. Subsequently, mediation by gestational weight gain was assessed Diabetologia (2019) 62:915–925 919 Table 1 Maternal and neonatal characteristics per intervention group Characteristic UC HE&PA HE PA Neonatal LTFU Maternala Age, years 31.9 ± 5.6 32.5 ± 5.3 32.4 ± 5.5 31.7 ± 4.9 31.5 ± 5.5 Multiparous 35 (44) 42 (53) 51 (55) 38 (46) 47 (46) European descent 71 (89) 69 (86) 79 (86) 68 (83) 91 (89) Living with partner 76 (95) 74 (93) 90 (98) 77 (94) 93 (91) Higher education 45 (56) 43 (54) 55 (60) 46 (56) 50 (49) Maternal smoking 13 (16) 7 (9) 17 (19) 13 (16) 17 (17) Paternal smoking 24 (30) 22 (28) 37 (40) 27 (33) 33 (33) Pre-pregnancy BMI, kg/m2 33.7 ± 3.7 33.6 ± 3.6 34.2 ± 4.6 33.8 ± 3.9 33.4 ± 3.9 Gestational weight gain at 35–37 weeks, kg 8.6 ± 4.6 6.4 ± 3.9b 7.6 ± 4.9 8.2 ± 4.9 8.3 ± 4.4 Gestational diabetes at 24–28 weeks 8 (10) 11 (14) 13 (14) 9 (11) 12 (16) Insulin or metformin after 24–28 weeks 1 (1) 3 (4) 1 (1) 0 (0) 3 (3) Neonatalc Female sex 45 (56) 41 (51) 41 (45) 39 (48) 33 (52) Gestational age at birth, weeks 39.7 ± 1.4 39.8 ± 1.4 39.6 ± 1.7 39.6 ± 1.4 38.9 ± 4.2 Birthweight, g 3581 ± 510 3480 ± 504 3485 ± 629 3469 ± 512 3399 ± 546 Birthweight <2500 g 4 (5) 1 (1) 5 (6) 2 (3) 4 (7) Birthweight >4000 g 18 (23) 15 (19) 18 (20) 13 (16) 7 (11) Data are n (%) or means ± SD a Maternal groups: UC, n=80; HE&PA, n=80; HE, n=92; PA, n=82; neonatal LTFU, n=102 (n=100 for paternal smoking; n=73 for gestational diabetes) b Gestational weight gain in the HE&PA group was the only characteristic different from the UC group (p=0.002) c Neonatal groups: UC, n=80; HE&PA, n=80; HE, n=91; PA, n=81; neonatal LTFU, n=64 (n=62 for birthweight <2500 g and >4000 g) LTFU, loss to follow-up using a simple mediation model. Sex dependency of all pathways flank (−0.5 mm, p = 0.03) skinfold thicknesses, sum of in the mediation models was tested. Models with the sum of skinfolds (−1.8 mm, p = 0.03), as well as a reduced estimated skinfolds or estimated fat percentage as outcome were adjusted fat mass (−63 g, p = 0.04) and estimated fat percentage for the time of measurement (in h) after birth. Models with ges- (−1.2%, p = 0.04) but no effect on estimated fat-free mass, tational weight gain as mediator were adjusted for the number of head or abdominal circumferences compared with UC days between measurements. (Table 3). Furthermore, cord blood leptin levels were reduced in this intervention group (−3.80 μg/l, p = 0.03). In the HE group, no significant differences were observed compared with UC. In the PA group, no differences in neonatal anthro- Results pometry were observed. However, a small but statistically significant reduction in cord blood leptin was found in female Figure 1 shows the randomisation and follow-up of participants neonates only (−5.79 μg/l, p = 0.05). Results from the sensi- in the trial. A total of 436 women were included in the trial. tivity analysis, using complete data, showed similar interven- Participants lost to follow-up (n = 102) were not different from tion effects on neonatal anthropometry (electronic supplemen- those for whom neonatal data were available (n = 334) (Table 1). tary material [ESM] Table 1). Maternal characteristics were comparable between intervention groups (Table 1), except that gestational weight gain was lower in the HE&PA group compared with the UC group. Mediation effects of lifestyle and gestational weight gain on neonatal anthropometry and cord blood Effects of interventions on neonatal anthropometry leptin and cord blood leptin In contrast to the multilevel analyses for intervention effects, Neonatal anthropometry and cord blood leptin per interven- mediation analyses did not detect significant interaction with tion group are described in Table 2. The HE&PA intervention neonatal sex (data not shown); therefore, sexes were com- was associated with smaller thigh (−0.7 mm, p = 0.04) and bined in all further mediation analyses. 920 Diabetologia (2019) 62:915–925 Table 2 Neonatal anthropometry and cord blood leptin according to intervention group Variable UC HE&PA HE PA n Mean±SD n Mean±SD n Mean±SD n Mean±SD Anthropometry Triceps skinfold, mm 69 5.3 ± 1.3 70 5.0 ± 1.3 74 5.3 ± 1.5 69 5.3 ± 1.6 Subscapular skinfold, mm 71 5.0 ± 1.4 70 4.6 ± 1.1 74 5.0 ± 1.4 68 4.9 ± 1.4 Thigh skinfold, mm 70 6.9 ± 1.9 69 6.2 ± 1.7 75 6.6 ± 2.1 68 6.5 ± 2.1 Flank skinfold, mm 70 4.5 ± 1.7 69 3.9 ± 1.1 75 4.4 ± 1.6 68 4.3 ± 1.3 Sum of skinfolds, mm 69 21.6 ± 5.2 69 19.8 ± 4.3 73 21.2 ± 5.8 68 21.0 ± 5.4 Head circumference, cm 78 34.8 ± 1.7 79 34.8 ± 1.5 90 34.6 ± 2.0 80 34.8 ± 1.9 Abdominal circumference, cma 73 33.4 ± 2.6 70 33.0 ± 2.7 76 33.3 ± 3.2 74 32.9 ± 2.7 Male 31 33.6 ± 2.7 33 32.7 ± 2.9 43 33.4 ± 3.5 37 33.2 ± 2.5 Female 42 33.3 ± 2.6 37 33.3 ± 2.5 33 33.2 ± 2.8 36 32.6 ± 3.0 Upper-arm circumference, cm 73 11.4 ± 1.2 71 11.2 ± 1.0 77 11.5 ± 1.3 71 11.5 ± 1.3 Lower-arm circumference, cm 73 10.0 ± 1.0 70 9.9 ± 1.0 77 10.0 ± 1.0 71 9.9 ± 1.0 Thigh circumference, cm 73 15.4 ± 1.7 71 15.0 ± 2.1 77 15.4 ± 1.8 71 15.3 ± 1.7 Calf circumference, cm 73 11.6 ± 1.2 70 11.5 ± 1.2 77 11.6 ± 1.2 71 11.6 ± 1.1 Estimated fat mass, g 67 511 ± 181 68 451 ± 171 69 492 ± 213 65 463 ± 190 Estimated fat percentage, % 67 13.8 ± 3.5 68 12.6 ± 3.4 69 13.2 ± 4.18 65 12.9 ± 3.8 Estimated fat-free mass, g 67 3111 ± 339 68 3029 ± 351 69 3105 ± 410 65 3021 ± 3547 Cord blood Leptin, μg/la 64 12.19 ± 13.01 50 8.29 ± 6.33 70 10.00 ± 7.95 57 9.10 ± 6.88 Male 31 7.57 ± 4.94 25 6.80 ± 4.93 37 8.21 ± 4.80 29 8.31 ± 7.37 Female 33 16.52 ± 16.52 25 9.77 ± 7.27 33 12.02 ± 10.12 28 9.91 ± 6.36 Differences between groups are shown in Table 3 a Data on male and female neonates are provided separately for variables for which an interaction of intervention effect with sex was found (p<0.10) First, we examined whether changes in lifestyle mediated percentage and cord blood leptin (ESM Table 5). No media- the intervention effects in the HE&PA and/or PA group on tion by gestational weight gain at 24–28 or at 35–37 weeks sum of skinfolds, estimated fat percentage and leptin (ESM was found. Gestational weight gain at 24–28 weeks was not Tables 2–4). In parallel mediation models, including all six significantly associated with leptin (β = 0.01; 95% CI −0.44, lifestyle factors, no mediation was found for sum of skinfolds 0.47) (Fig. 2b). The reductions in gestational weight gain at or estimated fat percentage (ESM Tables 2, 3). Sedentary be- 24–28 weeks did not mediate the effects of the HE&PA or PA haviour was the only lifestyle factor mediating intervention interventions on leptin (indirect effect HE&PA −0.01; 95% CI effects on leptin (ESM Table 4). Mediation of the intervention −0.47, 0.43; indirect effect PA 0.01; 95% CI −0.26, 0.42), and effect on leptin by sedentary behaviour was further analysed the direct associations of the interventions with leptin in a simple mediation model. From baseline to 24–28 weeks, remained significant (Fig. 2b). Sedentary behaviour at 24– sedentary behaviour was reduced in both the HE&PA and PA 28 weeks was not associated with gestational weight gain at interventions (Fig. 2a), and cord blood leptin was increased 24–28 (β = −0.02; 95% CI −0.06, 0.03) or at 35–37 weeks among women who had more sedentary behaviour (β = 0.37; (β = 0.01; 95% CI −0.06, 0.08). 95% CI 0.20, 0.54). The reduction in sedentary behaviour mediated the effects of the HE&PA and PA interventions on leptin (indirect effect HE&PA −0.96; 95% CI −2.74, −0.11; Discussion indirect effect PA −1.01; 95% CI −2.89, −0.11); consequently, the interventions were no longer directly associated with lep- Associations between maternal lifestyle and neonatal adipos- tin in an independent manner (Fig. 2a). Changes in sedentary ity exist [30–32], but causality has not been demonstrated. behaviour from baseline to 35–37 weeks did not mediate in- Here we used an RCT design to demonstrate the consequences tervention effects on leptin (ESM Table 4). of changing lifestyle in pregnancy on neonatal adiposity. The Second, we tested whether gestational weight gain mediat- combined HE&PA intervention resulted in a reduction of ed intervention effects on the sum of skinfolds, estimated fat thigh, flank skinfold thickness, sum of skinfolds, estimated Diabetologia (2019) 62:915–925 921 Table 3 Differences in neonatal outcomes per intervention group compared with UC group Neonatal outcome HE&PA vs UC HE vs UC PA vs UC β (95% CI) p value β (95% CI) p value β (95% CI) p value Anthropometrya Triceps skinfold, mm −0.3 (−0.8, 0.1) 0.14 −0.2 (−0.6, 0.3) 0.49 −0.1 (−0.5, 0.4) 0.74 Subscapular skinfold, mm −0.4 (−0.8, 0.1) 0.10 −0.1 (−0.5, 0.3) 0.65 −0.2 (−0.7, 0.2) 0.28 Thigh skinfold, mm −0.7 (−1.3, −0.03) 0.04 −0.4 (−1.1, 0.2) 0.17 −0.5 (−1.2, 0.1) 0.11 Flank skinfold, mm −0.5 (−0.9, −0.1) 0.03 −0.2 (−0.6, 0.2) 0.38 −0.2 (−0.6, 0.3) 0.49 Sum of skinfolds, mm −1.8 (−3.5, −0.2) 0.03 −0.8 (−2.5, 0.8) 0.33 −1.0 (−2.6, 0.7) 0.25 Head circumference, cm 0.0 (−0.5, 0.6) 0.99 −0.1 (−0.6, 0.5) 0.78 0.0 (−0.5, 0.6) 0.93 Abdominal circumference, cm −0.3 (−1.3, 0.5) 0.45 −0.2 (−1.0, 0.7) 0.72 −0.8 (−1.7, 0.1) 0.08 Male −1.3 (−2.7, 0.02)† 0.05 −0.3 (−1.6, 1.0) 0.63 −1.0 (−2.3, 0.3) 0.14 Female 0.3 (−0.9, 1.5)† 0.64 0.1 (−1.2, 1.3) 0.89 −0.9 (−2.1, 0.4) 0.17 Upper-arm circumference, cm −0.2 (−0.6, 0.2) 0.31 0.1 (−0.3, 0.4) 0.79 0.0 (−0.4, 0.4) 0.87 Lower-arm circumference, cm −0.2 (−0.5, 0.1) 0.26 0.0 (−0.3, 0.3) 1.00 −0.2 (−0.5, 0.2) 0.38 Thigh circumference, cm −0.5 (−1.0, 0.1) 0.13 −0.1 (−0.7, 0.5) 0.76 −0.2 (−0.8, 0.4) 0.44 Calf circumference, cm −0.1 (−0.5, 0.2) 0.49 −0.0 (−0.4, 0.4) 0.98 −0.1 (−0.5, 0.3) 0.63 Estimated fat mass, g −63 (−124, −2) 0.04 −23 (−84, 38) 0.45 −58 (−119, 4) 0.07 Estimated fat percentage, % −1.2 (−2.4, −0.04) 0.04 −0.7 (−1.8, 0.5) 0.27 −1.0 (−2.2, 0.2) 0.09 Estimated fat-free mass, g −78 (−198, 42) 0.20 −2 (−122, 118) 0.98 −82 (−203, 40) 0.19 Cord blood Leptin, μg/l −3.80 (−7.15, −0.45) 0.03 −2.01 (−5.09, 1.07) 0.20 −2.90 (−6.15, 0.34) 0.08 Male −0.77 (−3.74, 2.19) 0.61 0.64 (−2.05, 3.33) 0.64 0.74 (−2.11, 3.59)† 0.61 Female −6.50 (−12.31, −0.69) 0.03 −3.83 (−9.21, 1.55) 0.16 −5.79 (−11.43, −0.14)† 0.05 p values for the intervention effect are based on multilevel regression models, with the individual and site as levels a Analyses with neonatal anthropometry outcomes were adjusted for time (in h) after birth of the measurement. Further adjusting of analyses by parity, pre-pregnancy BMI, education, smoking, neonatal sex, gestational age at birth and mode of delivery did not change the results † p<0.10 for interaction with neonatal sex fat mass, fat percentage and cord blood leptin. There was no The clinical relevance of the reduction in neonatal adipos- effect on lean mass. In the PA intervention, a reduction in ity for future child health is unclear. Longitudinal studies dem- leptin was found in female neonates. Effects on leptin in both onstrated an association between lower neonatal adiposity and intervention groups were mediated through a reduction in sed- subsequent lower adiposity in infancy [36] and childhood [3, entary behaviour. Changes in gestational weight gain alone 37, 38]. The long-term follow-up of the offspring of the DALI did not mediate these intervention effects. lifestyle trial will help assess the relevance of our intervention The DALI lifestyle trial is the first comprehensive study to effects on the neonate for future obesity and health. demonstrate beneficial changes in neonatal anthropometric mea- The mediation analyses suggest that gestational weight sures, indicating a reduction in neonatal adiposity, a finding val- gain does not mediate intervention effects on neonatal out- idated by a reduction in cord blood leptin. No reduction in fat- comes. This finding differs from previous longitudinal cohort free mass was found, which was previously suggested as a pos- studies that reported associations of gestational weight gain sible risk factor for future cardiovascular and metabolic disease with neonatal adiposity [8, 39, 40]. This may be due, in part, to [33, 34]. Previous large RCTs of combined HE&PA interven- residual confounding in observational studies that was exclud- tions reported positive changes in physical activity and diet [16, ed by our RCT design. Furthermore, it might be that only early 17] but not in neonatal body composition [13, 14] or cord blood pregnancy gestational weight gain—obviously not influenced leptin [14]. A low glycaemic diet intervention resulted in a lower by our interventions—is related to neonatal adiposity [39], thigh circumference but did not affect any other neonatal body although such associations in later pregnancy have been re- composition outcome [35]. None of these previous interventions ported [8]. The most likely explanation, however, is that our focused on reducing sedentary behaviour, whereas DALI study included only obese women, and excessive gestational HE&PA and PA interventions highlighted the importance of re- weight gain has no or a much smaller effect on neonatal fat ducing sitting time during counselling sessions [18]. mass in this group [41, 42]. 922 Diabetologia (2019) 62:915–925 Fig. 2 Schematic diagram of the a Indirect effect results of simple mediation by a1×b HE&PA = −0.96 (−2.74, −0.11)* sedentary behaviour at 24–28 a2×b HE = −0.68 (−2.14, 0.03) a3×b PA = −1.01 (−2.89, −0.11)* weeks (n=239) (a) and gestational weight gain (GWG) at 24–28 weeks (n=238) (b) on cord blood Sedentary behaviour leptin. The path coefficients (MET h/week) b = 0.37 (0.20, 0.54)* between intervention condition, a1 HE&PA = −2.62 (−5.19, −0.06)* 24–28 weeks mediators (sedentary behaviour in a2 HE = −1.86 (−4.26, 0.53) a and GWG in b) and cord blood a3 PA = −2.76 (−5.29, −0.23)* leptin are shown. Statistically significant path coefficients (p<0.05) are indicated with Intervention condition Cord blood leptin asterisks and with solid lines. (μg/l) Direct effect Non-significant paths are c1‘ HE&PA = −3.25 (−6.61, 0.12) indicated with dashed lines. c2‘ HE = −1.76 (−4.88, 1.37) c3‘ PA = −2.55 (−5.87, 0.77) Indirect effects are calculated as the product of the coefficients of the (a) and (b) paths (a×b). MET, b Indirect effect metabolic equivalent of task a1×b HE&PA = −0.01 (−0.47, 0.43) a2×b HE = −0.004 (−0.28, 0.21) a3×b PA = 0.01 (−0.26, 0.42) GWG (kg) a1 HE&PA = −0.97 (−1.97, 0.04) 24–28 weeks b = 0.01 (−0.44, 0.47) a2 HE = −0.32 (−1.27, 0.62) a3 PA = 0.59 (−0.40, 1.58) Cord blood leptin Intervention condition (μg/l) Direct effect c1‘ HE&PA = −4.19 (−7.69, −0.70)* c2‘ HE = −2.42 (−5.66, 0.83) c3‘ PA = −3.65 (−7.07, −0.22)* The metabolic pathway(s) linking changes in sedentary pregnant women throughout Europe. A further strength is behaviour and neonatal adiposity remain(s) to be identified. the use of leptin as a biomarker of neonatal adiposity. Maternal glucose associates with risk of large-for-gestational- Although also produced by the placenta, cord blood leptin age neonates [43], but the DALI HE&PA and PA interven- levels correlate with neonatal adiposity (Spearman’s corre- tions had no effect on fasting and post-load glucose levels lation coefficient r = 0.46 with estimated fat mass in this [20]. However, a reduction in sedentary time was associated study [data not shown]) [21]. Leptin levels reflect total with reduced cord blood leptin levels. In previous observa- fetal fat mass and not only subcutaneous fat, in contrast tional studies, more sedentary time was associated with in- to skinfold thickness measurements. However, more direct creased maternal leptin [44], lipid levels [44, 45] and C- measures of neonatal adiposity, such as air displacement reactive protein levels [45] but not with insulin resistance or plethysmography, would have been preferred. The use of glucose [44, 46]. Reduced inflammation might be involved in self-reported data on lifestyle factors in the mediation anal- the pathway between sedentary behaviour and offspring adi- yses could be considered a limitation of our study. posity, a hypothesis worth testing. Therefore, it is possible that the counselling intervention Some strengths and limitations of our study need con- a woman received influenced her answers to questions on sidering. A strength of the pan-European DALI lifestyle lifestyle. We did not predefine a main outcome for the trial is its unique design in assessing effects of counselling analyses presented in this paper, which may be regarded interventions on healthy eating and physical activity sepa- as a limitation. A further limitation is the limited power of rately or in combination. Although data were missing for the study. Despite this, we found significant intervention some neonates, women providing data on neonatal anthro- effects on several outcomes that are highly related to each pometry or cord blood leptin were similar to the total study other, which adds confidence to the findings observed. sample. Furthermore, intervention effects were confirmed However, it does have consequences for the sample size in sensitivity analysis using complete data. Our results are of a future study, which should be adequately powered in therefore applicable to white, obese (BMI ≥29 kg/m2) order to be able to replicate/validate our findings. Diabetologia (2019) 62:915–925 923 Intervention effects on neonatal adiposity were due to reduced Contribution statement All authors contributed to the conception and/or design of the trial, read and corrected draft versions of the manuscript and sedentary behaviour but not increased moderate-to-vigorous approved the final manuscript. MNMvP undertook the statistical analy- physical activity. This might imply that increasing time spent in ses, wrote the first draft of the paper and is the guarantor of this work. light-intensity physical activity is important and is likely easier for obese women to implement in their daily lives. Importantly, Open Access This article is distributed under the terms of the Creative current guidelines for physical activity in pregnancy focus rec- Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, ommendations on maintaining sufficient levels of moderate-to- distribution, and reproduction in any medium, provided you give appro- vigorous physical activity [47, 48]. Should our present findings priate credit to the original author(s) and the source, provide a link to the be confirmed for other populations across the BMI and ethnicity Creative Commons license, and indicate if changes were made. ranges, guidelines on physical activity in pregnancy will need modification to include a reduction in sedentary behaviour. Since many recent lifestyle interventions in pregnancy, includ- Publisher’s note Springer Nature remains neutral with regard to jurisdic- tional claims in published maps and institutional affiliations. ing the DALI lifestyle trial, have failed to show beneficial effects on maternal glucose metabolism and incidence of gestational diabetes, it has been suggested that future efforts might instead be best invested in pre-conception lifestyle interventions. References However, our results show that whilst gestational diabetes risk 1. UNICEF, WHO, World Bank Group (2015) Levels and trends in was not reduced, lifestyle intervention in pregnancy is relevant child malnutrition: UNICEF–WHO–World Bank Group joint child for reducing neonatal adiposity and may have implications for malnutrition estimates. Available from https://data.unicef.org/wp- future childhood obesity. The lack of intervention effects on content/uploads/2015/12/JME-2015-edition-Sept-2015_203.pdf. birthweight also emphasises that birthweight is a crude measure Accessed 30 October 2018 of neonatal adiposity and demonstrates the importance of mea- 2. Spalding KL, Arner E, Westermark PO et al (2008) Dynamics of fat cell turnover in humans. Nature 453(7196):783–787. https://doi. suring body composition instead. org/10.1038/nature06902 In summary, the combined HE&PA intervention resulted in 3. McDermid E (2017) HAPO FUS shows persistent ill effects of less subcutaneous fat in neonates and lower cord blood leptin m a t e r n a l d i a b e t e s . Av a i l a b l e f r o m h t t p s : / / d i a b e t e s . levels. These effects were not mediated by gestational weight medicinematters.com/pregnancy/gestational-diabetes/hapo-fus- shows-persistent-ill-effects-of-maternal-diabetes/12358110. gain, despite a substantial reduction of 2.0 kg in the HE&PA Accessed 16 Jul 2018 group [20]. However, reduced sedentary behaviour mediated 4. Geserick M, Vogel M, Gausche R et al (2018) Acceleration of BMI the intervention effect on leptin. The mechanisms linking sed- in Early Childhood and Risk of Sustained Obesity. N Engl J Med entary behaviour and neonatal adiposity, and the implications 379(14):1303–1312. https://doi.org/10.1056/NEJMoa1803527 for future child obesity, need to be elucidated. 5. Goldstein RF, Abell SK, Ranasinha S et al (2017) Association of gestational weight gain with maternal and infant outcomes: a sys- Acknowledgements Open access funding provided by University of tematic review and meta-analysis. JAMA 317(21):2207–2225. Graz. We thank the participants, coaches, research midwives/nurses and https://doi.org/10.1001/jama.2017.3635 health professionals who collaborated in the recruitment. Some of the data 6. Gillman MW, Ludwig DS (2013) How early should obesity pre- were presented as an abstract at the ADA 78th Scientific Sessions in vention start? N Engl J Med 369(23):2173–2175. https://doi.org/10. 2018. 1056/NEJMp1310577 7. Sewell MF, Huston-Presley L, Super DM, Catalano P (2006) Data availability Data are available from the corresponding author on Increased neonatal fat mass, not lean body mass, is associated with reasonable request. maternal obesity. Am J Obstet Gynecol 195(4):1100–1103. https:// doi.org/10.1016/j.ajog.2006.06.014 Funding The project described has received funding from the European 8. Starling AP, Brinton JT, Glueck DH et al (2015) Associations of Community’s 7th Framework Programme (FP7/2007-2013) under grant maternal BMI and gestational weight gain with neonatal adiposity agreement no. 242187. In the Netherlands, additional funding was pro- in the Healthy Start study. Am J Clin Nutr 101(2):302–309. https:// vided by the Netherlands Organisation for Health Research and doi.org/10.3945/ajcn.114.094946 Development (ZonMw) (grant no. 200310013). In Poland, additional 9. Crowther CA, Hiller JE, Moss JR et al (2005) Effect of treatment of funding was obtained from the Polish Ministry of Science (grant no. gestational diabetes mellitus on pregnancy outcomes. N Engl J Med 2203/7, PR/2011/2). In Denmark, additional funding was provided by 352(24):2477–2486. https://doi.org/10.1056/NEJMoa042973 Odense University Free Research Fund. In the UK, the DALI team ac- 10. Landon MB, Spong CY, Thom E et al (2009) Multicenter, random- knowledges the support received from the National Institute for Health ized trial of treatment for mild gestational diabetes. N Engl J Med Research Clinical Research Network Eastern, especially the local diabe- 361(14):1339–1348. https://doi.org/10.1056/NEJMoa0902430 tes clinical and research teams based in Cambridge. In Spain, additional 11. Gillman MW, Oakey H, Baghurst PA, Volkmer RE, Robinson JS, funding was provided by CAIBER (Consorcio de Apoyo a la Crowther CA (2010) Effect of treatment of gestational diabetes Investigación Biomédica en Red; 1527-B-226). The funders had no role mellitus on obesity in the next generation. Diabetes Care 33(5): in any aspect of the study beyond funding. 964–968. https://doi.org/10.2337/dc09-1810 12. Landon MB, Rice MM, Varner MW et al (2015) Mild gestational Duality of interest The authors declare that there is no duality of interest diabetes mellitus and long-term child health. Diabetes Care 38(3): associated with this manuscript. 445–452. https://doi.org/10.2337/dc14-2159 924 Diabetologia (2019) 62:915–925 13. Dodd JM, Deussen AR, Mohamad I et al (2016) The effect of 28. Hayes AF (2013) Introduction to mediation, moderation, and con- antenatal lifestyle advice for women who are overweight or obese ditional process analysis: a regression-based approach. Guilford, on secondary measures of neonatal body composition: the LIMIT New York randomised trial. BJOG 123(2):244–253. https://doi.org/10.1111/ 29. Hayes AF, Preacher KJ (2014) Statistical mediation analysis with a 1471-0528.13796 multicategorical independent variable. Br J Math Stat Psychol 14. Poston L, Bell R, Briley AL et al (2017) Improving pregnancy outcome 67(3):451–470. https://doi.org/10.1111/bmsp.12028 in obese women: the UK Pregnancies Better Eating and Activity 30. Hayes L, Bell R, Robson S, Poston L, UPBEAT Consortium (2014) randomised controlled Trial. NIHR Journals Library, Southampton Association between physical activity in obese pregnant women 15. Brown J, Alwan NA, West J et al (2017) Lifestyle interventions for and pregnancy outcomes: the UPBEAT pilot study. Ann Nutr the treatment of women with gestational diabetes. Cochrane Metab 64(3–4):239–246. https://doi.org/10.1159/000365027 Database Syst Rev 5:CD011970 31. Horan MK, McGowan CA, Gibney ER, Donnelly JM, McAuliffe 16. Dodd JM, Turnbull D, AJ MP et al (2014) Antenatal lifestyle advice FM (2014) Maternal low glycaemic index diet, fat intake and post- for women who are overweight or obese: LIMIT randomised trial. prandial glucose influences neonatal adiposity—secondary analysis BMJ 348:g1285. https://doi.org/10.1136/bmj.g1285 from the ROLO study. Nutr J 13(1):78. https://doi.org/10.1186/ 17. Poston L, Bell R, Croker H et al (2015) Effect of a behavioural 1475-2891-13-78 intervention in obese pregnant women (the UPBEAT study): a 32. Bisson M, Tremblay F, St-Onge O, Oostdam N et al (2017) multicentre, randomised controlled trial. Lancet Diabetes Influence of maternal physical activity on infant's body composi- Endocrinol 3(10):767–777. https://doi.org/10.1016/S2213- tion. Pediatr Obes 12(Suppl S1):38–46. https://doi.org/10.1111/ 8587(15)00227-2 ijpo.12174 18. Jelsma JG, van Poppel MN, Galjaard S et al (2013) DALI: vitamin 33. Catalano PM, Mele L, Landon MB et al (2014) Inadequate weight D and lifestyle intervention for gestational diabetes mellitus (GDM) gain in overweight and obese pregnant women: what is the effect on prevention: an European multicentre, randomised trial—study pro- fetal growth? Am J Obstet Gynecol 211(2):137.e1–137.e7. https:// tocol. BMC Pregnancy Childbirth 13(1):142. https://doi.org/10. doi.org/10.1016/j.ajog.2014.02.004 1186/1471-2393-13-142 34. Barker DJ, Osmond C, Golding J, Kuh D, Wadsworth ME (1989) 19. Simmons D, Jelsma JG, Galjaard S et al (2015) Results from a Growth in utero, blood pressure in childhood and adult life, and European multicenter randomized trial of physical activity and/or mortality from cardiovascular disease. BMJ 298(6673):564–567. healthy eating to reduce the risk of gestational diabetes mellitus: the https://doi.org/10.1136/bmj.298.6673.564 DALI Lifestyle Pilot. Diabetes Care 38(9):1650–1656. https://doi. 35. Donnelly JM, Walsh JM, Byrne J, Molloy EJ, McAuliffe FM org/10.2337/dc15-0360 (2015) Impact of maternal diet on neonatal anthropometry: a ran- 20. Simmons D, Devlieger R, van Assche A et al (2017) Effect of domized controlled trial. Pediatr Obes 10(1):52–56. https://doi.org/ physical activity and/or healthy eating on GDM risk: the DALI 10.1111/j.2047-6310.2013.00216.x lifestyle study. J Clin Endocrinol Metab 102(3):903–913 36. Sauder KA, Kaar JL, Starling AP, Ringham BM, Glueck DH, 21. Hauguel-de Mouzon S, Lepercq, Catalano P (2006) The known and Dabelea D (2017) Predictors of infant body composition at 5 unknown of leptin in pregnancy. Am J Obstet Gynecol 194(6): months of age: the Healthy Start Study. J Pediatr 183:94–9.e1. 1537–1545. https://doi.org/10.1016/j.ajog.2005.06.064 https://doi.org/10.1016/j.jpeds.2017.01.014 22. World Health Organization (2013) Diagnostic criteria and classifi- 37. Simpson J, Smith AD, Fraser A et al (2017) Programming of adi- cation of hyperglycaemia first detected in pregnancy. WHO/NMH/ posity in childhood and adolescence: associations with birth weight MND/13.2. World Health Organization, Geneva and cord blood adipokines. J Clin Endocrinol Metab 102(2):499– 23. Jelsma JGM, Simmons D, Gobat N et al (2017) Is a motivational 506. https://doi.org/10.1210/jc.2016-2342 interviewing based lifestyle intervention for obese pregnant women 38. Catalano PM, Farrell K, Thomas A et al (2009) Perinatal risk factors across Europe implemented as planned? Process evaluation of the for childhood obesity and metabolic dysregulation. Am J Clin Nutr DALI study. BMC Pregnancy Childbirth 17(1):293. https://doi.org/ 90(5):1303–1313. https://doi.org/10.3945/ajcn.2008.27416 10.1186/s12884-017-1471-9 39. Josefson JL, Simons H, Zeiss DM, Metzger BE (2016) Excessive 24. Deierlein A, Thornton J, Hull H, Gallagher D (2012) An anthropo- gestational weight gain in the first trimester among women with metric model to estimate neonatal fat mass using air displacement normal glucose tolerance and resulting neonatal adiposity. J plethysmography. Nutr Metab 9(1):21. https://doi.org/10.1186/ Perinatol 36(12):1034–1038. https://doi.org/10.1038/jp.2016.145 1743-7075-9-21 40. Crozier SR, Inskip HM, Godfrey KM et al (2010) Weight gain in 25. Chasan-Taber L, Schmidt MD, Roberts DE, Hosmer D, Markenson G, pregnancy and childhood body composition: findings from the Freedson PS (2004) Development and validation of a Pregnancy Southampton Women’s Survey. Am J Clin Nutr 91(6):1745– Physical Activity Questionnaire. Med Sci Sports Exerc 36(10):1750– 1751. https://doi.org/10.3945/ajcn.2009.29128 1760. https://doi.org/10.1249/01.MSS.0000142303.49306.0D 41. Hull HR, Thornton JC, Ji Y et al (2011) Higher infant body fat with 26. Simmons D, Mandell C, Fleming C, Gatland B, Leakehe L (1994) excessive gestational weight gain in overweight women. Am J Evaluation of a diabetes knowledge and behaviour (DKB) ques- Obstet Gynecol 205(3):211.e1–211.e7. https://doi.org/10.1016/j. tionnaire. Asia Pac J Clin Nutr 3(4):193–200 ajog.2011.04.004 27. Jelsma JG, van Leeuwen KM (2016) Beliefs, barriers, and prefer- 42. Waters TP, Huston-Presley L, Catalano PM (2012) Neonatal body ences of European overweight women to adopt a healthier lifestyle composition according to the revised institute of medicine recom- in pregnancy to minimize risk of developing gestational diabetes mendations for maternal weight gain. J Clin Endocrinol Metab mellitus: an explorative study. J Pregnancy 2016:3435791 97(10):3648–3654. https://doi.org/10.1210/jc.2012-1781 Diabetologia (2019) 62:915–925 925 43. HAPO Study Cooperative Research Group, Metzger BE, Lowe LP in pregnant and non-pregnant women. BMC Pregnancy Childbirth et al (2008) Hyperglycemia and adverse pregnancy outcomes. N 11(1):44. https://doi.org/10.1186/1471-2393-11-44 Engl J Med 358:1991–2002 47. Perales M, Artal R, Lucia A (2017) Exercise during pregnancy. JAMA 44. Nayak M, Peinhaupt M, Heinemann A et al (2016) Sedentary be- 317(11):1113–1114. https://doi.org/10.1001/jama.2017.0593 havior in obese pregnant women is associated with inflammatory 48. Evenson KR, Barakat R, Brown WJ et al (2014) Guidelines for markers and lipid profile but not with glucose metabolism. physical activity during pregnancy: comparisons from around the Cytokine 88:91–98. https://doi.org/10.1016/j.cyto.2016.08.031 world. Am J Lifestyle Med 8(2):102–121. https://doi.org/10.1177/ 45. Loprinzi PD, Fitzgerald EM, Woekel E, Cardinal BJ (2013) 1559827613498204 Association of physical activity and sedentary behavior with bio- logical markers among U.S. pregnant women. J Women's Health (Larchmt) 22(11):953–958. https://doi.org/10.1089/jwh.2013.4394 46. Gradmark A, Pomeroy J, Renstrom F et al (2011) Physical activity, sedentary behaviors, and estimated insulin sensitivity and secretion Affiliations Mireille N. M. van Poppel 1,2 & David Simmons 3,4 & Roland Devlieger 5,6 & F. Andre van Assche 5,6 & Goele Jans 5,6 & Sander Galjaard 5,6,7 & Rosa Corcoy 8,9 & Juan M. Adelantado 8 & Fidelma Dunne 10,11 & Jürgen Harreiter 12 & Alexandra Kautzky-Willer 12 & Peter Damm 13 & Elisabeth R. Mathiesen 13 & Dorte M. Jensen 14,15,16 & Lise-Lotte Andersen 14,15,16 & Mette Tanvig 14,15,16,17 & Annunziata Lapolla 18 & Maria G. Dalfra 18 & Alessandra Bertolotto 19 & Ewa Wender-Ozegowska 20 & Agnieszka Zawiejska 20 & David Hill 21 & Frank J. Snoek 22 & Judith G. M. Jelsma 1 & Gernot Desoye 23 1 12 Department of Public and Occupational Health, Amsterdam Public Gender Medicine Unit, Endocrinology and Metabolism, Health Research Institute, Amsterdam UMC, Vrije Universiteit Department of Internal Medicine III, Medical University of Vienna, Amsterdam, Amsterdam, the Netherlands Vienna, Austria 2 13 Institute of Sport Science, University of Graz, Mozartgasse 14, Center for Pregnant Women with Diabetes, Departments of 8010 Graz, Austria Endocrinology and Obstetrics, Rigshospitalet, Institute of Clinical 3 Medicine, Faculty of Health and Medical Sciences, University of Macarthur Clinical School, Western Sydney University, Copenhagen, Copenhagen, Denmark Sydney, NSW, Australia 14 4 Steno Diabetes Center Odense, Odense University Hospital, Institute of Metabolic Science, Addenbrooke’s Hospital, Odense, Denmark Cambridge, UK 15 5 Department of Gynaecology and Obstetrics, Odense University Department of Development and Regeneration: Pregnancy, Fetus Hospital, Odense, Denmark and Neonate, KU Leuven, Leuven, Belgium 16 6 Department of Clinical Research, Faculty of Health Sciences, Department of Gynaecology and Obstetrics, University Hospitals University of Southern Denmark, Odense, Denmark Leuven, Leuven, Belgium 17 7 Region of Southern Denmark, Denmark Division of Obstetrics and Prenatal Medicine, Department of 18 Obstetrics and Gynaecology, Erasmus University Medical Center, Dipartimento di Medicina, Università Degli Studi di Padova, Rotterdam, the Netherlands Padua, Italy 8 19 Institut de Recerca de l’Hospital de la Santa Creu i Sant Pau, Department of Clinical and Experimental Medicine, Azienda Barcelona, Spain Ospedaliero Universitaria Pisana, Pisa, Italy 9 20 Centro de Investigación Biomédica en Red (CIBER) Medical Faculty I, Poznań University of Medical Sciences, Bioengineering, Biomaterials and Nanotechnology, Instituto de Poznań, Poland Salud Carlos III (ISCIII), Madrid, Spain 21 Recherche en Santé Lawson SA, Bronschhofen, Switzerland 10 Galway Diabetes Research Centre, National University of Ireland, 22 Department of Medical Psychology, Amsterdam Public Health Galway, Ireland Research Institute, Amsterdam UMC, Vrije Universiteit 11 College of Medicine, Nursing and Health Sciences, National Amsterdam, Amsterdam, the Netherlands University of Ireland, Galway, Ireland 23 Department of Obstetrics and Gynaecology, Medical University of Graz, Graz, Austria