Gestational Diabetes Risk and Low Birth Weight After Metabolic Bariatric Surgery: a Complex Interplay to be Balanced

In parallel to the rising prevalence of obesity in the general population, the prevalence of obesity in women during pregnancy is also increasing, which is estimated to range between 7 and 25% in Europe [1, 2]. Women with obesity are at increased risk of short- and long-term pregnancy complications for both mother and kindred [3]. Adverse obesity-related pregnancy outcomes include gestational diabetes (GD), hypertensive disorders of pregnancy (HDP), prematurity, cesarean delivery, perinatal mortality, and congenital malformations [4].

Given that metabolic bariatric surgery (MBS) is being increasingly performed in women at reproductive age, pregnancies after MBS are also becoming progressively more frequent [5, 6]. The available evidence shows that MBS is associated with a significant reduction in the risk of several obesity-associated diseases and most particularly GD [7]. However, concerns have also been raised on the potential harms, such as fetal growth restriction (FGR) and small-for-gestational-age (SGA) infants [8, 9]. In a recent meta-analysis, pregnant women after MBS matched for preconception (PC) BMI to non-operated controls were shown to be at increased risk of SGA infants and preterm delivery, suggesting that despite the benefits, surgery also carries additional risks. However, given that the risk for adverse obstetric outcomes was higher than observed in women matched for PC BMI, suggests that these could be independent of BMI [8]. Notwithstanding, data interpretation is hampered by the heterogeneity of populations, as most of the original evidence available on MBS and pregnancy outcomes derives from case–control studies [10] and the fact that controls were not always comparable regarding body mass index (BMI) and maternal age [11]. Consequently, it is not entirely clear whether the benefits, as well as the risks on maternal and neonatal outcomes, are solely the end result of weight loss or whether there is also a contributing role of surgery, itself. Therefore, aim was to compare the obstetric and neonatal outcomes of women previously submitted to MBS to those of non-operated women matched for age and PC BMI.

Women submitted to MBS prior to conception were identified among those admitted for antenatal care during the pre-specified time period (n = 91). After applying the exclusion criteria, 12 pregnancies were excluded for the following reasons: delivery at another institution (n = 9), pregnancy after assisted reproduction techniques (n = 2) and twin pregnancy (n = 1), yielding the final number of eligible pregnancies for analysis (n = 79), which were pair matched to controls (n = 79) as above described. Women who underwent MBS clinical features before surgery and conception are displayed at Table 1. Besides age and PC BMI, operated and non-operated women were also matched for other non-specified clinical features and despite smoking before and/or during pregnancy being more frequent in the MBS group, the difference was not statistically significant (26.6% vs. 15.2%; p = 0.08) (Table 2).

Significant differences in the pre-specified pregnancy outcomes were identified between groups—namely GD was less frequent in the MBS group (7.6% vs. 19%; p = 0.03). First and second trimester fasting glucose, birth weight and newborn height were significantly lower in MBS group, while the frequency of SGA (22.8% vs. 6.3%; p < 0.01) was higher in MBS group compared to controls (Table 3). NICU admission occurred in four newborns of the MBS group (severe fetal growth restriction (n = 1), renal malformation with pulmonary dysplasia (n = 1), neonatal bradycardia (n = 1) and respiratory distress (n = 1)) and in a newborn of the control group (due to prematurity) (5.1% vs. 1.3%; p = 0.37). No neonatal deaths occurred. Maternal complications during puerperium were more frequent in the MBS group (17.7% vs. 5.1%; p < 0.01), being anemia and wound infections the most accountable entities (p = 0.01 and p = 0.03 respectively) (Table 3). There were no significant differences between groups regarding gestational age at delivery, induction of labor or mode of delivery.

On regression analysis, pre-operative BMI was not found to be significantly associated with any of the pregnancy outcomes (supplementary Table 1), while smoking was the only variable found to be associated with a higher risk of SGA (supplementary Table 2). Hence, a subanalysis excluding pregnancies from smoking mothers was then performed to disentangle the influence of this variable on pregnancy outcomes. From this subanalysis, fasting glucose at 24–28 gestational weeks, birth weight and birth height remained significantly lower; however, the SGA frequency was no longer statistically different in the MBS group when compared to control group (15.5% vs. 6%; p = 0.14) (Table 4). A sub-group analysis comparing pregnancies of women who conceived < 12 months vs > 12 month after MBS to evaluate the potential impact on birth weight and the occurrence of SGA was performed, but no significant differences were found between the groups (supplementary Table 3).

MBS is a highly effective weight loss intervention and weight loss decreases the risk of adverse obstetric outcomes in women with obesity. However, to what extent is MBS able to revert the risk of obesity-associated obstetric complications towards the levels of age and BMI matched women, without incurring at additional surgical-related risks, is not entirely known [3, 7]. Therefore, to address this knowledge gap, we decided to compare the pregnancy outcomes of women previously submitted to MBS to those of non-operated women matched for age and PC BMI.

Our study shows that GD was significantly less frequent, while birth weight was significantly lower after MBS when compared to age and PC BMI matched controls, after controlling for smoking.

Our case–control study matched for PC BMI, which contrasts to previous studies that used pre-surgery BMI as control [14], allowed us to further explore and disentangle the effects of MBS at reducing the incidence of GD beyond the weight loss achieved. At first glance, our study findings corroborate the previously available evidence on GD after MBS [7, 15, 16]. Furthermore, despite there were no differences regarding PC BMI and age, our finding of a lower fasting glucose, which exerts a major influence over GD risk, supports the hypothesis that during pregnancy MBS also influences glycemic dynamics in a body weight independent manner.

Nevertheless, the so far available evidence on GD and MBS must be interpreted with caution, as the majority of the studies used a 75 g or 100 g OGTT for GD diagnosis [17], despite the known caveat of depicting altered glucose kinetics, with early hyperglycemia leading to an increased rate of false positive GD diagnoses [18]. Indeed, fasting glucose levels were previously shown to be lower in women after MBS when compared both to lean or BMI-matched controls, while the postprandial glucose excursion curve showed a characteristic pattern, including a glucose rise at 60 min followed by reactive hypoglycemia, which occurred in 54.8% of the cases [19]. Although there is no doubt that alternative GD diagnostic criteria are still needed for this specific population [18], in the meantime, frequent capillary blood glucose self-monitoring is the diagnostic resource most frequently used [20]. The fact that in our study GD screening relied on capillary blood glucose measurements further strengthens the finding of a reduced prevalence of GD after MBS, irrespectively of the weight loss achieved or GWG.

MBS was also associated with a higher prevalence of SGA and reduced newborn’s birth weight and height, despite there were no significant differences on GWG or time from surgery to conception. Despite the difficulties to differentiate a SGA from FGR fetus [21], the association between MBS and low birth weight, regardless of the matching performed, seems to be fairly unanimous [9, 22‐24]. Kwong et al. found that the risk of SGA in women after BMS was higher when compared to women matched for pre-surgery BMI, as well as, PC BMI with an odds ratio (OR) of 2.16 and an OR of 2.23, respectively [8]. In a subsequent study by Jacamon et al., MBS was associated with a reduced risk of excessive fetal growth with a trend for a higher incidence of SGA, despite matching on PC BMI [15]. Another study concluded that in singleton gestations, women with obesity who underwent a prior MBS, had a higher risk of FGR, which further supports the association of MBS with FGR, regardless the weight loss or BMI achieved [25]. The lack of essential nutrients in result of a combination of food restriction and malabsorption was hypothesized to be responsible for the negative impact on fetal growth [26]. A short surgery-to-conception interval and a GWG lower than recommended by the Institute of Medicine (IOM) are additional factors that may influence fetal growth after MBS [14, 23]. Moreover, a greater glycemic variability, a lower glucose nadir during the OGTT and a greater incidence of postprandial hypoglycemia were also proposed to have a negative impact on the development of fetuses from women previously submitted to gastric bypass [27]. Nevertheless, detailed data on glucose kinetics during pregnancies after different BMS procedures are not currently available [19]. Continuous glucose monitoring during pregnancy after BMS could not only prove to be a valid alternative to diagnose GD, but also provide detailed information on glucose kinetics, most particularly the occurrence of asymptomatic hypoglycemic events, and how these impact on pregnancy outcomes [28]. Moreover, this data could be useful to guide clinicians and dieticians to provide targeted macronutrient intake counseling to avoid hypoglycemic events, potentially harmful to the fetus [29].

When managing pregnancy in women after MBS it is important to take into account several factors, obesity related or not, which may influence the obstetric outcomes [30, 31]. Smoking is recognized to be a major risk factor for lower fetal weight, irrespective of mother’s surgical status [32]. After excluding smoking mothers from the pregnancies analysis, there was no longer a significant difference in SGA, although both fasting glucose mean and birth weight remained significantly lower.

Daily low-dose ASA is part of standard obstetric care to reduce the risk for preeclampsia, preterm birth, small for gestational age/intrauterine growth restriction, and perinatal mortality in pregnant women at high risk for preeclampsia [33]. Pregnant women submitted to MBS frequently fulfil criteria for ASA prescription (previous history of preeclampsia, BMI greater than 30, maternal age greater or equal to 35 years) [34]. As a consequence, the use of ASA was more prevalent in women submitted to surgery.

Maternal anemia during gestation and puerperium was also more frequent in pregnancies of women who had undergone MBS. In our study, less than half of the women previously submitted to MBS reported to adhere to vitamin supplement recommendations prior to conception, and more than one in ten women conceived within one year after the surgery despite medical counselling to postpone pregnancy plans until after the second postoperative year. Although these women were under regular follow-up with a multidisciplinary obesity team, it has been clearly shown that most patients either fail to perceive the potential benefits of vitamin supplementation compliance or have financial constraints, which render micronutrient deficiencies a frequent complication after MBS [11],

The main strengths of this study are the adjustment for two major variables for the outcomes being studied (age and PC BMI), the fact that PC BMI and not pre-surgery BMI was considered, and that all women were followed up at a single center by the same multidisciplinary team comprising obstetricians, endocrinologists and dietitians. The GD screening was performed by frequent capillary glucose self-monitoring; data on fasting glucose was included in the analysis, and the impact of smoking was considered on obstetrics outcomes. However, there are also limitations that must be acknowledged. First, the study retrospective design and small sample size are likely to have hindered the statistical significance of some parameters. Second, not all women underwent MBS at the same institution nor by the same surgical team. Third, given the small numbers, it was not possible to evaluate a potential influence of different types of MBS procedures on the study outcomes. Forth, it was not possible to evaluate to what extent the outcomes found were due to the PC BMI rather than residual from the pre-surgery BMI. Finally, although smoking arises as an important variable accounting for the occurrence of SGA, we were unable to disentangle the relative contribution of active vs inactive smoking due to lack of data.

MBS was associated with a prevalence of GD that was lower than observed in non-operated women with the same age and BMI. After controlling for smoking habits, this occurred at the expense of a lower birth weight but without criteria for SGA. Our data reinforces the hypothesis that MBS has body weight independent effects on glucose kinetics during pregnancy with potentially distinctive impacts for mother and offspring, which need to be balanced.