The BAMBI II Study

Background Glucose metabolism in pregnancy following bariatric surgery Bariatric surgery (BS) increases fertility and decreases obesity related risks associated with pregnancy such as macrosomia, gestational diabetes mellitus (GDM), and pre-eclampsia (1, 2, 3). However, BS is also associated with alterations to the glucose metabolism (4). These alterations have been shown to be more pronounced following Roux-en-Y gastric bypass (RYGB) compared to sleeve gastrectomy (SG) given that RYGB causes more extensive modifications to the gastro-intestinal system (5). Meals are associated with an exaggerated secretion of insulin, which can cause a mismatch between glucose and insulin levels with a risk of hypoglycemia, also known as post-bariatric hypoglycemia (4, 5). On the other hand, pregnancy is accompanied by a progressively decreasing insulin sensitivity (6), which could counteract the risk of hypoglycemia. Results from the first Bariatric surgery And consequences for Mother and Baby In pregnancy (BAMBI) study showed that hypoglycemia was present in both early and late pregnancy as measured by continuous glucose monitoring (CGM) and mixed meal test (7, 8). Surprisingly, the BAMBI study revealed a two-fold increase in time spent in hyperglycemia during CGM throughout pregnancy and in the postpartum period among pregnant women with RYGB compared to BMI-matched healthy pregnant controls (7). Therefore, identification of women with GDM is important. Oral glucose tolerance testing (OGTT) is not recommended among women with bariatric surgery given the risk of hypoglycemia following the glucose load as well as the lack of validated cut-offs (9). According to our current national guideline, diagnosis of GDM in post-bariatric pregnant women is based on self-monitoring of fasting glucose and glucose 1.5 hour following meals (10). However, as shown for both non-pregnant and pregnant individuals with previous bariatric surgery, nadir glucose appears at about 1.5 hour following meals (5, 8). Thus, we might underestimate the prevalence of gestational diabetes in this population. Given the risks associated with GDM for both the mother and the offspring (11), correctly identifying women with GDM is crucial.

Please see attached figure 1, a schematic overview of different outcomes in pregnancy following bariatric surgery.

Fetal growth During the initial 12 months post-surgery, the body is in a catabolic state characterized by a significant weight loss including both lean and fat body mass (12). As of now, it is recommended to postpone conception at least 12 months after BS in order to reduce the risk of insufficient gestational weight gain (GWG) and thereby restricted fetal growth (9). Small-for-gestational age (SGA) neonates are increased in prevalence following BS (13). Insufficient GWG have been associated with an increased risk of delivering SGA neonates, yet insufficient GWG only account for about half the cases (13, 14). The BAMBI study showed a tendency towards increased time spent in hypoglycemia and decreased time in hyperglycemia among the women with RYGB delivering SGA neonates as compared to the women with RYGB delivering appropriate-for-gestational age (AGA) neonates. Additionally, women with RYGB delivering large-for-gestational (LGA) neonates spent less time in hypoglycemia and more time in hyperglycemia as compared to the women with RYGB delivering AGA neonates (8). Unfortunately, the BAMBI study lacked power to show a statistical significant difference.

In Denmark, female infertility and BMI \>35kg/m2 serve as eligibility criteria for BS (15). About 40% of people having BS are women of childbearing age, both nationally and internationally (16, 17). The number of bariatric surgeries performed among women of childbearing age have been increasing since 2015, and 890 women had bariatric surgery in Denmark in 2022 (16). Therefore, studies comparing RYGB and SG are needed to guide the women to the best possible pregnancy outcome when having BS before pregnancy.

Aims

In women with BS we aim to study:

1. Trimester-specific occurrence of hypo- and hyperglycemia 2. Association between glucose excursions, fetal growth and neonatal body composition

Hypotheses

1. Pregnant women with BS spend more time in hyperglycemia (blood glucose \>7.8 mmol/l) than women without BS, which is associated with higher birthweight and higher neonatal fat mass. Furthermore, women with RYGB spend more time in hyperglycemia than women with SG. 2. Pregnant women with BS have a higher frequency of both symptomatic and asymptomatic hypoglycemia (blood glucose \<3.0 mmol/l) than women without BS, and this is associated with lower birthweightand lower neonatal fat mass. Additionally, RYGB induces a higher risk of hypoglycemia and larger glucose variability than SG.

Methods Study design Longitudinal, prospective, multicenter cohort study. Participants Study participants will be enrolled in early pregnancy at the Departments of Gynecology and Obstetrics at Esbjerg and Grindsted Hospital, Lillebaelt Hospital, Odense University Hospital, Gødstrup Hospital, and Aarhus University Hospital.

Inclusion criteria

* RYGB group: pregnant women with RYGB and their neonate (n = 75) * SG group: pregnant women with SG and their neonate (n = 75) * non-BS group: pregnant women without BS matched on age, pre-pregnancy-BMI and parity and their neonate (n = 75)

Exclusion criteria

* multiple pregnancy * age below 18 and above 45 years * ongoing smoking or substance abuse * severe psychiatric disorder or chronic disease * use of GLP-1 analogue within two months of conception * diabetes

* women with overt diabetes at inclusion (HbA1c ≥ 48 mmol/l and/or fasting p-glucose ≥ 7 mmol/l) * women with pre-gestational diabetes (type1 or 2) prior to RYGB * women with GDM in a previous pregnancy will not be excluded During a two-year period we aim to include 75 pregnant women with RYGB, 75 pregnant women with SG, 75 age-, BMI- and parity-matched controls without BS.

Maternal clinical information on date and type of BS, weight trajectories (before BS, pre-gestational, gestational, post-partum), post-operative and pregnancy related complications, and blood pressure will be collected by the primary investigator from hospital records.

GDM in women with previous SG or RYGB will be diagnosed by self-monitoring of glucose profiles according to current national guidelines.

A 2-hour 75 g OGTT will be performed in the control group in gestational week 24 to test for GDM.

Glucose metabolism Glucose levels during everyday life will be assessed using the Dexcom G6 sensor in each trimester and 4-6 weeks post-partum. The Dexcom G6 sensor lasts for 10 days and does not require calibration. Self-Monitoring of Blood Glucose (SMBG) will be used for confirmation of hypoglycemia. Hypoglycemia will be defined as blood glucose \<3.5 mmol/L (low blood glucose) and \< 3.0 mmol/L (very low blood glucose). During the entire study period, the women will report any hypoglycemic symptoms with concomitant SMBG.

Blood samples measuring maternal insulin resistance (HOMA-IR) and vitamin status (hemoglobin, ferritin, vitamin B12, vitamin D) will be drawn in early-, mid-, and late pregnancy as well as 8 weeks postpartum. At birth, cord blood will be drawn in order to measure neonatal insulin resistance (HOMA-IR).

Fetal growth Antenatal serial ultrasound measurements will be performed at gestational week 28 and 34.

Measurements will include abdominal circumference (AC) and calculation of estimated fetal weight deviation.

Postnatally, anthropometric measures including birthweight and length, AC, skinfold measurements and body composition estimated by the Pea Pod system within 72 hours after birth will be performed. The Pea Pod is an air displacement plethysmography system using whole body densitometry to determine body composition (fat and fat-free mass) in infants.

Lifestyle At inclusion, the women will be equipped with an activity tracker. The activity tracker will register daily number of steps as well as exercises performed during pregnancy and in the postpartum period. During the 10 days CGM measurements, the participants will register dietary intake. At each study visit, the participants will fill out questionnaires on diet, physical activity, physical and mental health (Edinburg Postnatal Depression Scale (EDPS), Pregnancy Physical Activity Questionnaire (PPAQ-DK), Food Frequency World Health Organization - Five Well-Being Index (WHO-5), Pittsburgh Sleep Quality Index (PSQI)). After delivery, questionnaires on breastfeeding will be added.

Outcomes

Primary outcomes:

• Time spent with interstitial glucose below 3.0 and 3.5 mmol/L and above 7.8 mmol/L

Secondary outcomes:

* Number of neonates born SGA * Number of neonates born LGA * Birth weight z-score * Body composition in the new-born

Investigation diagram Please see attached figure 2 of the BAMBI II study design.

Clinical impact and novelty The BAMBI II study addresses the United Nation's Sustainable Development Goal number three "Good health and well-being". Due to an increase in the number of SG operations performed, the number of women of reproductive age having BS is on the rise (16). As of now, the evidence on the subject is sparse. A better understanding of maternal metabolism and fetal growth in women with previous BS will support risk stratification and development of treatments both before and during pregnancy. This study will help clinicians to guide women of reproductive age considering BS and improve the care of pregnant women with BS as well as their offspring.

Work plan and time schedule Please see attached figure 3.

Statistics and power calculation We will evaluate changes from first trimester to postpartum visit within groups (paired testing). Furthermore, crude comparisons on single-measure outcomes between groups will be done applying unpaired Student's t test or Wilcoxon-Mann-Whitney U test, while trajectories (repeated measures) between groups will be done by applying repeated measured ANOVA models. Linear regression models will be used to assess the differences between groups at different levels of adjustment for relevant confounders.

Sample size calculations on Continuous Glucose Monitoring (CGM) outcome:

In the BAMBI study, pregnant women with RYGB spend 0.6-1.1% (SD 0.4-2.3%) of the time with glucose \< 3.0 mmol/L compared to 0-0.2% (SD 0-0.5%) among BMI-matched controls. Similar figures for time spent with glucose \> 7.8 mmol/l was 9.0-11.6% (SD 5.1-6.7%) and 3.6-6.6% (SD 3.9-8.8%).

We did crude comparisons on single-measure outcomes between groups for each of the trimesters and the postpartum period with unpaired students t-test and the Wilcoxon ranksum test. The power calculation for this part of the BAMBI II study is based on the main outcome: time in hypo- and hyperglycemia during CGM, which in BAMBI II will be compared for each of the measuring points. As such, these data will be handled independently of each other. In order to calculate the sample size for hypo- and hyperglycemia individually in each trimester and the postpartum period, 13 to 45 participants should be included in each group.

Sample size calculations on fetal growth (birthweight percentile) outcome:

In the BAMBI study increased time spent in hypoglycemia was defined as more than 4% of the time spent with interstitial glucose below 3.5 mmol/L according to an international consensus on the treatment of pregnant women with pregestational diabetes. The birthweight percentile for women with Roux-en-Y gastric bypass (RYGB) and increased time in hypoglycemia was 25 (SD 28) compared to 51 (SD 35) among those with appropriate time in hypoglycemia. To show a difference between these 2 subgroups we would need a sample of 52 women with RYGB.

To allow for a 10% dropout, we aim to include 75 pregnant women in each group. Limitations and main challenges

* The 10 days CGM and questionnaires 3 times in pre