Associations of dietary glycemic index and load during pregnancy with blood pressure, placental hemodynamic parameters and the risk of gestational hypertensive disorders

Purpose The aim of this study was to examine the associations of dietary glycemic index and load with gestational blood pressure, placental hemodynamic parameters and the risk of gestational hypertensive disorders. Methods In a population-based cohort among 3378 pregnant Dutch women, dietary glycemic index and load were assessed from food frequency questionnaires at median 13.4 (95% range 9.9–22.9) weeks gestation. Blood pressure was measured in early-, mid- and late-pregnancy. Placental hemodynamic parameters were measured in mid- and late-pregnancy by ultrasound. Data on gestational hypertensive disorders was acquired from medical records. Results Mean dietary glycemic index (SD) was 58 (3) and mean dietary glycemic load (SD) was 155 (47). Dietary glycemic index was not associated with blood pressure, placental hemodynamic parameters and the risk of gestational hypertensive disorders. Higher dietary glycemic load SDS was associated with a higher diastolic blood pressure in early-pregnancy, remaining after adjustment for socio-demographic and lifestyle factors ((0.98 (95% CI 0.35–1.61) mmHg per SDS increase in glycemic load). No other associations of glycemic load with blood pressure or placental hemodynamic parameters and the risk of gestational hypertensive disorders were present. No significant associations of dietary glycemic index and load quartiles with longitudinal blood pressure patterns from early to late-pregnancy were present. Conclusion Within this low-risk pregnant population, we did not find consistent associations of dietary glycemic index and load with blood pressure, placental hemodynamic parameters and the risk of gestational hypertensive disorders. Further studies need to assess whether the effects on gestational hemodynamic adaptations are more pronounced among high-risk women with an impaired glucose metabolism. Supplementary Information The online version contains supplementary material available at 10.1007/s00394-021-02670-5.


INDEX SUPPLEMENTAL MATERIAL Page
3 Table S1 Non-response analysis: characteristics of the total study population versus participating women without data on dietary intake 4 Figure S1 Directed Acyclic Graph for initial confounder selection.
5 Table S2 Longitudinal associations of glycemic index and load with systolic and diastolic blood pressure a 6 Table S3 A) Basic models: Associations of dietary glycemic index and load quartiles with systolic and diastolic blood pressure during pregnancy in total population B) Confounder models: Associations of dietary glycemic index and load quartiles with systolic and diastolic blood pressure during pregnancy in total population 8 Table S4 A) Basic models: Associations of dietary glycemic index and load quartiles with umbilical artery pulsatility index, uterine artery resistance index and bilateral notching in total population B) Confounder models: Associations of dietary glycemic index and load quartiles with umbilical artery pulsatility index, uterine artery resistance index and bilateral notching in total population 10 Table S5 A) Basic models: Associations of dietary glycemic index and load quartiles with hypertensive disorder of pregnancy, gestational hypertension and preeclampsia in total population B) Confounder models: Associations of dietary glycemic index and load quartiles with hypertensive disorder of pregnancy, gestational hypertension and preeclampsia in total population 12 Table S6 A) Sensitivity analysis: Associations of dietary glycemic index with clinical cutoff for low-and normal-glycemic diet with systolic and diastolic blood pressure during pregnancy B) Sensitivity analysis: Associations of dietary glycemic index with clinical cutoff for low-and normal-glycemic diet with umbilical artery pulsatility index, uterine artery resistance index and bilateral notching C) Sensitivity analysis: Associations of dietary glycemic index with clinical cutoff for low and normal glycemic diet with hypertensive disorder of pregnancy, gestational hypertension and preeclampsia Table S7 A) Sensitivity analysis: Associations of dietary glycemic index and load with systolic and diastolic blood pressure during pregnancy in population with BMI≥25 B) Sensitivity analysis: Associations of dietary glycemic index and load with umbilical artery pulsatility index, uterine artery resistance index and bilateral notching in population with BMI≥25 C) Sensitivity analysis: Associations of dietary glycemic index and load with hypertensive disorder of pregnancy, gestational hypertension and preeclampsia in population with BMI≥25 18 Table S8 A) Sensitivity analysis: Associations of dietary glycemic index and load with systolic and diastolic blood pressure during pregnancy in population with study enrollment <14 weeks of gestation B) Sensitivity analysis: Associations of dietary glycemic index and load with uterine artery resistance index, umbilical artery pulsatility index and bilateral notching in population with study enrollment <14 weeks of gestation C) Sensitivity analysis: Associations of dietary glycemic index and load with hypertensive disorder of pregnancy, gestational hypertension and preeclampsia in population with study enrollment <14 weeks of gestation 21  Figure S2. Directed Acyclic Graph for confounder selection.
Supplementary Table S2. Longitudinal associations of glycemic index and load with systolic and diastolic blood pressure from repeated measurement models a CI, Confidence interval. a Values are based on repeated non-linear regression models and reflect the change in blood pressure in mmHg per glycemic index and glycemic load quartile compared to women with the highest dietary quality (quartile 1) as reference. Models are adjusted for gestational age at the time of intake. b P-value reflects the significance level of the estimate. CI, Confidence Interval. GI, glycemic index. GL, glycemic load. a Values are regression coefficients (95% confidence interval) and reflect the difference in mmHg blood pressure per glycemic index or glycemic load quartile. Groups are compared to women with the highest dietary quality (quartile 1) as reference. Estimates are from multiple imputed data. b Models are adjusted for gestational age at time of intake. c Tests for trend were based on multiple linear regression models with glycemic index and load as z-score. *P-value <0.05.  1.41 (0.77, 2.60) ncases=18 UmPI, umbilical artery pulsatility index. UtRI, uterine artery resistance index. CI, Confidence Interval. GI, glycemic index. a Values are regression coefficients (95% confidence interval) and reflect the differences in umbilical artery pulsatility index and uterine artery resistance index of women with a low-glycemic index diet compared to women with a normal-glycemic index diet as reference. Estimates are from multiple imputed data. b Values are odds ratios (95% confidence interval) from multiple logistic regression models and reflect the differences in risks of bilateral notching of women with a normal-glycemic index diet compared to women with a low-glycemic index diet as reference. Estimates are from multiple imputed data. c Basic models are adjusted for gestational age at time of intake. d Confounder models are adjusted for maternal age, educational level, parity, prepregnancy BMI, kcal, smoking habits, alcohol use, folic acid use and gestational age at time of the measurements. *P-value <0.05.

Supplementary table 6c.
Sensitivity analysis: Associations of dietary glycemic index with clinical cut-off for low and normal glycemic diet with hypertensive disorder of pregnancy, gestational hypertension and preeclampsia (n=3,298) a .
OR, odds ratio. CI, Confidence Interval. GI, glycemic index. a Values are odds ratios (95% confidence interval) from multiple logistic regression models and reflect the difference in risks of gestational hypertensive disorders, gestational hypertension and preeclampsia of women with a low-glycemic index diet compared to women with a normal-glycemic index diet as reference. Estimates are from multiple imputed data. b Basic models are adjusted for gestational age at time of intake. c Confounder models are adjusted for maternal age, educational level, parity, prepregnancy BMI, kcal, smoking habits, alcohol use, folic acid use and gestational age at time of the measurements. *P-value <0.05. .25) SDS, standard deviation score. CI, Confidence Interval. UmPI, umbilical artery pulsatility index. UtRI, uterine artery resistance index. a Values are regression coefficients (95% confidence interval) from multiple linear regression models and reflect the differences in umbilical artery pulsatility index and uterine artery resistance index per one increase in standard deviation score of maternal glycemic index and glycemic load. Estimates are from multiple imputed data. b Values are odds ratios (95% confidence interval) from multiple logistic regression models and reflect the difference in risks of bilateral notching per one increase in standard deviation score of maternal glycemic index and load. Estimates are from multiple imputed data. c Basic models are adjusted for gestational age at time of intake. d Socio-demographic models are adjusted for maternal age, educational level, parity and gestational age at time of measurements. e Lifestyle models are adjusted for maternal age, educational level, parity, prepregnancy BMI, kcal, smoking habits, alcohol use, folic acid use and gestational age at time of the measurements.*P-value <0.05.

Gestational hypertensive disorders
Supplementary table 7c. Sensitivity analysis: Associations of dietary glycemic index and load with hypertensive disorder of pregnancy, gestational hypertension and preeclampsia in population with BMI≥25 (n=766) a . SDS, standard deviation score. CI, Confidence Interval. b Values are odds ratios (95% confidence interval) from multiple logistic regression models and reflect the difference in risks of gestational hypertensive disorders, gestational hypertension and preeclampsia per one increase in standard deviation score of maternal glycemic index and glycemic load. Estimates are from multiple imputed data. c Basic models are adjusted for gestational age at time of intake. d Socio-demographic models are adjusted for maternal age, educational level, parity and gestational age at time of intake. e Lifestyle models are adjusted for maternal age, educational level, parity, prepregnancy BMI, kcal, smoking habits, alcohol use, folic acid use and gestational age at time of intake. *P-value <0.05.