Discussion
To the best of our knowledge, this is the first longitudinal characterization of the postprandial glucose metabolism and enteropancreatic hormone responses to a MMT in pregnancy following RYGB. The postprandial profiles of glucose and the enteropancreatic hormones did not change markedly during pregnancy. Compared with the controls, the women with RYGB were less insulin resistant, as evaluated by fasting insulin, the Stumvoll index, and HOMA2-IR throughout pregnancy. However, similar to the controls, they experienced an increase in fasting insulin and HOMA2-IR from the first to the third trimester. In spite of this pregnancy-induced increase in fasting-derived indices of insulin resistance, the frequency of biochemical hypoglycemia in women with RYGB was higher than in controls in both trimesters. Nadir plasma glucose was preceded by markedly elevated levels of total GLP-1 and insulin in women with RYGB, and these associations were significant in the first trimester but lost in the third trimester, where elevated plasma glucagon correlated with increased nadir glucose levels. Glucagon levels were lower in women with RYGB.
The profiles of glucose, insulin, C-peptide, total GLP-1, and total GIP among women with RYGB in the BAMBI cohort resembled the previously reported profiles in non-pregnant individuals treated with RYGB.2–4 15 Likewise, the prevalence of biochemical hypoglycemia corresponded to the prevalence found by Kefurt et al32 in a cohort consisting predominantly of women with a similar presurgery and postsurgery BMI, although slightly older and with a longer surgery-to-test interval. As in controls, insulin resistance increased in women with RYGB during pregnancy. However, this was not accompanied by a significantly lower frequency of biochemical hypoglycemia in the third trimester of pregnancy. Thus, even though the pregnancy-induced insulin resistance alters glucose metabolism from the first to the third trimester, these alterations do not seem to protect against the increased risk of postprandial biochemical hypoglycemia elicited by RYGB.
In this study, HOMA2-IR was lower in both the first and the third trimester of pregnancy among the women with RYGB compared with BMI-matched controls. This is in accordance with the findings by Svane et al33 showing lower HOMA-IR among individuals with RYGB compared with BMI-matched controls. They investigated individuals with RYGB without a history of type 2 diabetes and normal glucose tolerance. As both control groups were BMI-matched, RYGB could seem to improve the insulin sensitivity beyond the improvement caused by the weight loss. However, the reliability of HOMA as a surrogate measure of insulin sensitivity in individuals with RYGB has been questioned owing to increased insulin clearance, as was also evident in the current study.34 To reduce the influence of the increased insulin clearance, we based our HOMA2-IR calculation on fasting serum C-peptide. In addition, the lower degree of insulin resistance in the pregnant women with RYGB was confirmed by increased insulin sensitivity (Stumvoll index) during the MMT. In contrast, a recent prospective study found no differences in the metabolic benefits, including improved insulin sensitivity after an ~18% wt loss achieved by either RYGB or diet, suggesting that the metabolic benefits were related to weight loss itself.35 In that study, insulin sensitivity was measured using hyperinsulinemic euglycemic pancreatic clamp. Further studies are needed to establish if these differential findings are due to differences in study design, magnitude of weight loss after RYGB or methods of assessing insulin sensitivity.
As expected in normal pregnancy,9 fasting insulin, peak insulin, AUC of insulin, and HOMA2-IR increased during pregnancy among the controls. Insulin clearance was increased among the women with RYGB as previously reported in non-pregnant populations with RYGB,36 37 but did not change during pregnancy in either group. Therefore, an increase in HOMA2-IR could be interpreted as an increase in insulin resistance. During pregnancy, women with RYGB became more insulin resistant yet to a lesser extent than the controls.
Mean fasting plasma glucose decreased slightly during pregnancy among the women with RYGB. In individuals with type 2 diabetes, prediabetes, or obesity, mean fasting plasma glucose has been shown to decrease gradually after bariatric surgery.2 4 Ilesanmi et al2 showed that mean fasting plasma glucose decreased progressively from presurgery to 1-year postsurgery and stabilized between 1 and 2 years postsurgery in parallel with weight stabilization. Among the women with RYGB in the BAMBI study, 30 months had passed from surgery until pregnancy and inclusion, and the gestational weight gain was about 10 kg.10 Thus, the detected decrease in plasma glucose is not likely to be explained by the time lag of about 20–22 weeks between the two measurements.
Following RYGB, nutrients are delivered readily to the GLP-1 secreting L-cells in the terminal ileum and colon,38 and GLP-1 has been suggested as the main driver behind hypoglycemia during mid-pregnancy OGTT.39 In accordance, we found increased peak total GLP-1 levels among women with RYGB as compared with the controls in both trimesters. Furthermore, in the first trimester, we found that peak/AUC of serum insulin and plasma total GLP-1 levels correlated inversely with nadir plasma glucose after adjustment for potential confounders. Intriguingly, these correlations were lost in the third trimester, and replaced by significant correlation between early AUC of plasma glucagon and nadir plasma glucose levels. These results are in alignment with the study by Ilesanmi et al2 who found that peak GLP-1 and glucagon levels during the MMT were positively and negatively associated with increased time spent in hypoglycemia during CGM, respectively, in individuals with RYGB who had prediabetes or type 2 diabetes prior to RYGB.2 As proposed by Øhrstrøm et al,17 counterregulatory responses could modulate the risk of postbariatric biochemical hypoglycemia in addition to the increase in GLP-1. Our results indicate a shift in the role of elevated GLP-1/insulin versus lower glucagon in the risk of postbariatric biochemical hypoglycemia from the first to the third trimester in women with RYGB. Further studies are warranted to establish these findings and potential underlying mechanisms.
Even though a higher percentage of women with RYGB had biochemical hypoglycemia during the MMT compared with controls, the early postmeal response of plasma glucagon, 60 min AUC of glucagon, decreased among the women with RYGB while it was unchanged among the controls during pregnancy. The glucagon levels observed in the women with RYGB in the first trimester of our study are in line with the levels reported in previous studies of postbariatric individuals showing an early postprandial peak and the absence of an increase in plasma glucagon following glucose nadir.16–18 33 In the third trimester, both the early postprandial peak in plasma glucagon and the increase following glucose nadir appeared to be absent among the women with RYGB in our study. As speculated in the non-pregnant population, abnormal α-cell function could contribute to postbariatric biochemical hypoglycemia with deterioration in late pregnancy.18 40–42
GIP is secreted by the K-cells of the duodenum and proximal jejunum,43 and thus, most GIP-secreting cells are bypassed by the RYGB. In previous studies where GLP-1 was blocked, GIP failed to increase the postprandial insulin response to ingested glucose.38 44 Consequently, GIP is not believed to play an important role in postbariatric biochemical hypoglycemia.44 45 Similar to the findings by Leutner et al,39 who investigated pregnant women with RYGB by OGTT, we found equal responses of total GIP elicited by the MMT among women with RYGB and controls in both trimesters.
A strength of the study is the inclusion of a BMI-matched control group. The liquid meal was chosen in order to ensure ingestion of the same meal over the course of time. Recently, Hedbäck et al46 showed that liquid and solid meals with identical macronutrient composition result in similar postprandial glucose responses for both subjects with RYGB and non-surgical controls. They concluded that a liquid meal is suitable for the assessment of the glycemic and enteropancreatic response in both individuals with RYGB and non-surgical controls. Finally, the use of the new assay protocol of the Mercodia Glucagon ELISA is a strength. The new assay protocol reduces cross-reactivity with glicentin and proglucagon, which otherwise interferes with measurements particularly in individuals with RYGB, but makes comparisons with previous studies using other assay protocols difficult.
Our prespecified time points for blood sampling may limit our results regarding the exact determination of nadir plasma glucose levels and counterregulatory increases in plasma glucagon. Thus, both could happen later than 120 min. Jørgensen et al4 showed that the lowest glucose levels occur between 90 and 120 min in response to an MMT in normal glucose-tolerant individuals with RYGB, but later (180–240 min) in individuals with type 2 diabetes after RYGB. Apart from the two women with RYGB and one among the controls diagnosed with GDM, the women were perceived as normal glucose tolerant. Furthermore, Jørgensen et al4 did not report major changes in plasma glucagon after 180 min. Another potential limitation is the lack of measures of other glucose counterregulatory hormones. However, in a study of non-pregnant individuals with RYGB, cortisol and norepinephrine levels did not increase unless glucose was lowered by glucose-lowering drugs.17 Insulin resistance increases with age, and, therefore, the difference in age could potentially also affect our results.47 However, the women with RYGB were older, and thus, the reported differences in insulin resistance and biochemical hypoglycemia between the groups would be expected to be attenuated rather than overestimated. The correlation analyses were adjusted for age.
In conclusion, we demonstrated that pregnant women with RYGB had similar distinct profiles of postprandial glucose and the enteropancreatic hormones in response to a MMT in the first and the third trimester of pregnancy. These profiles resembled previously shown profiles for non-pregnant individuals with RYGB. Postbariatric biochemical hypoglycemia occurred in both the first and the third trimester of pregnancy despite an increase in insulin resistance during pregnancy. As such, pregnancy seems to neither aggravate nor ameliorate the risk of biochemical hypoglycemia as assessed by MMT. During the first trimester MMT, peak and AUC of total plasma GLP-1 and serum insulin levels were negatively associated with nadir plasma glucose, while these associations were absent in the third trimester and replaced by a positive association between the early postmeal plasma glucagon response and nadir plasma glucose. Thus, an increased understanding of the counterregulatory mechanisms, as illustrated by glucagon responses, could lead to prevention of hypoglycemia in this population.
Leave a Reply