This article is mentioned in:

Introduction

During normal pregnancy, numerous physiological and biochemical changes occur to meet the escalating need for nutrients to cover the energy requirement for the growing fetus. One of these changes is an increase in the release of the insulin hormone, known as hyperinsulinemia. This hyperinsulinemia arises as the gestational age advances, reaching peak levels in the third trimester (1). Therefore, it is widely accepted that pregnancy per se is a state of insulin resistance (IR) (1). IR is a condition characterized by impaired glucose uptake and utilization by peripheral tissues despite the presence of hyperinsulinemia (1). The exact cause of IR in normal pregnancy is not yet fully understood; yet, a number of risk factors have been identified, such as maternal obesity, physical inactivity, placentation, genetics and epigenetic factors (1).

IR is considered a precipitating factor for obstetrics-related disorders, such as gestational diabetes and preeclampsia (2-4). Nevertheless, pregnant women with IR may remain normotensive and euglycemic throughout their pregnancy (5). The known effects of hyperinsulinemia extend to the sympathetic nervous system, increasing the sympathetic tone and inducing renal sodium retention. This retention shifts sodium movement inside the cells, particularly in the smooth muscle of the blood vessels (6,7). These changes may raise blood pressure and precipitate hypertension (8).

There is a paucity of publications investigating IR in normal pregnant women. However, it has been reported that IR precedes the clinical emergence of hypertension by a few years in non-pregnant women (9). An association has also been observed between insulin hormone indices and blood pressure in healthy adolescents prior to the onset of hypertension (10).

Little is known about IR indices and blood pressure in normal pregnancies. Therefore, the present study was conducted in an aim to investigate the possible association between IR indices and blood pressure among healthy pregnant women screened negative for preeclampsia and gestational diabetes. The findings presented herein may guide clinicians in the early categorization of women who are at an increased risk of developing high blood pressure, preeclampsia and perhaps, gestational diabetes before becoming symptomatic.

Subjects and methods

The present cross-sectional study was conducted at the Antenatal Care Clinic of Saad Abuelela Maternity Hospital, Khartoum, Sudan, from February through December, 2021. Ethical clearance was obtained from the Department of Obstetrics and Gynecology, Faculty of Medicine, University of Khartoum (Khartoum, Sudan). All subjects included in the present study were briefed about its objectives and all subjects gave their written consent to participate.

Inclusion criteria

All the participants selected for the present study were healthy, pregnant Sudanese women aged ≥18 years with singleton pregnancies. Screening for gestational diabetes mellitus and IR occurred between the 24 and 28th weeks of gestation. For each woman, a well-structured questionnaire was filled out to gather sociodemographic, clinical, and medical history data.

Exclusion criteria

Candidates were excluded from the study if they had a previous history of hypertension, diabetes mellitus, thyroid disease, liver disease, kidney disease or severe anemia (hemoglobin levels <7 g/dl) or if they were smokers or on regular medication(s) that may affect blood pressure or blood glucose levels.

A total of 133 pregnant women were enrolled in the present study, for each of whom blood pressure was measured twice following a resting period of ~10 min using an OMRON3 automated blood-measuring device (OMRON Healthcare). There was a 1-to-2-min time interval between the two readings, and the mean of the two readings was then calculated. The reading was rejected in the case that the difference between the two readings was >5 mmHg, and the measurement was repeated until the readings became stable. The OMRON M3 readings can be validated among pregnant women, including among women with preeclampsia. The mean ± SD of differences between the OMRON 3 and the standard mercury sphygmomanometer was -1.6±2.8 mmHg for systolic blood pressure and -0.1±2.3 mmHg for diastolic blood pressure. This was rated and rated A/A for systolic and diastolic blood pressure (11). The calculations for body mass index (BMI) were based on the measurements of the height (meters) and weight (kg) of each candidate. BMI was calculated using the following equation: BMI=weight in kg/height in m2.

For each candidate, 2 ml venous blood sample was collected in a fluoride vacutainer in a fasting state (following overnight fasting for 10 h). A glucose tolerance test was then performed. Plasma glucose levels were measured by the enzymatic colorimetric method (BioSpectrometer, Eppendorf) using glucose oxidase reactions (Thermo Fisher Scientific, Inc.), following the manufacturer's instructions. Another sample of 2 ml blood was collected in ethylenediaminetetraacetic acid vacutainers to assess glycosylated hemoglobin (HbA1c). An i-chroma™ device (Biotech Med) was used to measure the HbA1c levels. The validation of the i-chroma™ device revealed a coefficient of variations of <2% in both the high and low HbA1c readings. The Spearman's correlation rank between the i-chroma device and the HPLC method is relatively low (Rho=0.368), and the mean bias is -0.50±1.62% (-5.5±17.7 mmol/mol) (12). Fasting levels of plasma insulin were measured using the immunoassay analyzer AIA 360 (Tosoh Bioscience). IR was calculated using the following homeostatic model assessment for IR (HOMA-IR) formula: Fasting insulin (U/l) x fasting glucose (mg/dl)/405(13). The quantitative insulin sensitivity check index (QUICKI) was calculated using the following formula: 1/(log (fasting insulin µU/ml) + log (fasting glucose mg/dl) (14). Homeostatic model assessment for β-cell function (HOMA-β) was measured based on fasting plasma glucose (FPG) and fasting plasma insulin (FPI) concentrations using the following formula: HOMA-β=FPI concentration (µU/ml) x20/FPG (mmol/l)-3.5(15).

Ethics approval and consent to participate

The present study received ethical clearance from the Department of Obstetrics and Gynaecology, Faculty of Medicine, University of Khartoum (No. 032, 2020). Signed informed consent was collected from all participants prior to enrolment.

Sample size calculation

The sample size was calculated using the Sample Size Calculators based on the significant minimum difference in the Pearson's correlations (r=0.30) between the systolic, diastolic blood pressure and HOMA-IR. The needed sample was 133, with 80% power and 5% precision at α=0.05.

Statistical analysis

Data were entered into a computer using SPSS (version 20.0) software (IBM Corp.). The Shapiro-Wilk test was used to assess the continuous variables for normality. All continuous variables were abnormally distributed. Accordingly, the median [interquartile range (IQR)] was used to express the abnormally distributed variables. Univariate linear regression analysis was performed with systolic and diastolic blood pressure as dependent variables, and maternal age, parity, education level, residency, job status, indices of IR and fasting insulin levels as independent variables. Variables with a P-value ≤0.200 in the univariate analysis were shifted to built-up multivariable linear regression analysis to assess the confounder. The coefficient with a 95% confidence interval (CI) and the P-value were reported. A two-sided P-value <0.05 was considered to indicate a statistically significant difference.

Results

Sociodemographic and clinical data

A total of 133 pregnant women were enrolled in the present study. The median (IQR) of age and parity were 28.0 years (24.0-32.0 years) and 1 (0.0-2.0), respectively. The majority (90.2%) of these women had a secondary level of education or higher. Of note, three quarters (75.9%) of the study participants were housewives, and 72.2% resided in urban residences. The observed range and median (IQR) of the systolic blood pressure of the participants was 100-130 mmHg, 110 mmHg (110-110.7 mmHg). For diastolic blood pressure, the observed range and median (IQR) was 60-80 mmHg, 70 mmHg (70.0-72.4 mmHg); for median blood pressure, the observed range and median (IQR) was 73.3-100 mmHg, 83 mmHg (83.0-85.1 mmHg) among the study participants. Fasting glucose levels ranged between 48-82 mg/dl with a median (IQR) of 70 mg/dl (63.0-78.0 mg/dl). Fasting insulin levels ranged between 0.5-12.5 mg/dl with a median (IQR) of 2.2 mg/dl (1.1-4.6 mg/dl). The median (range) was 0.356 (0.18-0.76) for HOMA-IR, 58.0 (9.4-144.0) for HOMA-β, and 0.463 (0.40-0.53) for QUICKI (Table I).

Table I Sociodemographic, clinical and biochemical characteristics of the pregnant women (n=133) included in the present study.

Table I

Sociodemographic, clinical and biochemical characteristics of the pregnant women (n=133) included in the present study.

Variables	Value
Education level, n (%)
≥Secondary	120 (90.2)
<Secondary	13 (9.8)
Residence, n (%)
Urban	96 (72.2)
Rural	37 (27.2)
Employment, n (%)
Housewife	101 (75.9)
Employed	32 (24.1)
Age (years), median (IQR)	28.0 (24.0-32.0)
Parity (number), median (IQR)	1.0 (0.0-2.0)
Systolic blood pressure (mmHg), median (IQR)	110 (110-110.7)
Diastolic blood pressure (mmHg), median (IQR)	70 (70.0-72.4)
Mean blood pressure (mmHg), median (IQR)	83 (83.0-85.1)
Fasting blood glucose (mg/dl), median (IQR)	70 (63.0-78.0)
Fasting insulin (µIU/ml), median (IQR)	2.2 (1.1-4.6)
Homeostatic model assessment for insulin resistance, median (IQR)	0.356 (0.18-0.76)
Quantitative insulin sensitivity check index, median (IQR)	0.463 (0.40-0.53)
Homeostatic model assessment for β-cell function, median (IQR)	58.0 (9.4-144.0)
Body mass index (kg/m2), median (IQR)	26.9 (23.9-30.2)

[i] IQR, interquartile range.

Univariate linear regression analysis

Univariate linear regression analysis revealed that fasting insulin levels (β=0.95; 95% CI, 0.10 to 1.80; P=0.027), QUICKI (β=-24.39; 95% CI, -43.1 to -5.59; P=0.012) and diastolic blood pressure (β=0.76; 95% CI, 0.44 to 1.08; P=<0.001) were associated with systolic blood pressure (Table II). Of note, parity (β=0.83; 95% CI, 0.07 to 1.58; P=0.031), fasting insulin levels (β=0.63; 95% CI, 0.15 to 1.10; P=0.010), HOMA-IR (β= 3.19; 95% CI, 0.67 to 5.71; P=0.013) and QUICKI (β=-11.97; 95% CI, -22.72 to -1.22; P=0.029) were the factors that were associated with diastolic blood pressure (Table III).

Table II Univariate and multivariate linear regression analysis of factors associated with systolic blood pressure in pregnant women (n=133) Khartoum, Sudan, 2021.

Table II

Univariate and multivariate linear regression analysis of factors associated with systolic blood pressure in pregnant women (n=133) Khartoum, Sudan, 2021.

	Univariate linear analysis			Multivariate linear analysis
Variables	Coefficient	95% Confidence interval	P-value	Coefficient	95% Confidence interval	P-value
Age, years	0.20	(-0.09 to 0.50)	0.170	0.18	(-0.16 to 0.53)	0.308
Parity	1.13	(-0.21 to 2.047)	0.098	0.67	(-0.90 to 2.25)	0.397
Education level	-3.61	(-8.61 to 1.37)	0.154	-4.48	(-9.42 to 0.45)	0.075
Residency	-0.12	(-4.64 to 4.39)	0.956
Employment	0.62	(-3.24 to 4.48)	0.751
Body mass index	0.29	(-0.01 to 0.60)	0.062	0.16	(-0.15 to 0.48)	0.312
Fasting blood glucose	-0.06	(-0.22 to 0.08)	0.398
Fasting insulin levelsa	0.95	(0.10 to 1.80)	0.027b	0.54	(-0.28 to 1.38)	0.195
Homeostatic model assessment for insulin resistancea	4.47	(-0.01 to 8.96)	0.050b	2.26	(0.32 to 0.99)	0.314
Quantitative insulin sensitivity check indexa	-24.39	(-43.1 to -5.59 )	0.012b	-16.11	(-35.55 to 3.32)	0.103
Homeostatic model assessment for β-cell functiona	0.10	(-0.01 to 0.06)	0.798
Diastolic blood pressure	0.76	(0.44 to 1.08)	<0.001b	0.64	(0.30 to 0.98)	<0.001b

[i] ^aThese factors were added one by one in the multivariable model. The multivariate model was adjusted for age, parity, education level, body mass index, fasting insulin, homeostatic model assessment for insulin resistance, quantitative insulin sensitivity check index and diastolic blood pressure.

[ii] ^bIndicates statistical significance (P<0.05).

Table III Univariate and multivariate linear regression analysis of factors associated with diastolic blood pressure in pregnant women (n=133), Khartoum, Sudan, 2021.

Table III

Univariate and multivariate linear regression analysis of factors associated with diastolic blood pressure in pregnant women (n=133), Khartoum, Sudan, 2021.

	Univariate linear analysis			Multivariable linear analysis
Variables	Coefficient	95% Confidence interval	P-value	Coefficient	95% Confidence interval	P-value
Age, years	0.16	(-0.01 to 0.35)	0.051	0.10	(-0.08 to 0.29)	0.247
Parity	0.83	(0.07 to 1.58)	0.031b	0.62	(-0.24 to 1.50)	0.158
Education level	0.13	(-2.72 to 2.99)	0.925
Residency	0.26	(-2.29 to 2.86)	0.837
Employment	-1.19	(-3.37 to 0.98)	0.279
Body mass index	0.157	(-0.01 to 0.33)	0.078	0.05	(-0.12 to 0.23)	0.562
Fasting blood glucose	-0.02	(-0.11 to 0.06)	0.627
Fasting insulin levelsa	0.63	(0.15 to 1.10)	0.010b	0.48	(0.02 to 0.94)	0.040b
Homeostatic model assessment for insulin resistancea	3.19	(0.67 to 5.71)	0.013b	2.51	(0.08 to 4.95)	0.043b
Quantitative insulin sensitivity check indexa	-11.97	(-22.72 to -1.22)	0.029b	-8.89	(-19.92 to 2.14)	0.113
Homeostatic model assessment for β-cell functiona	0.00	(-0.003 to 0.003)	0.861

[i] ^aThese factors were added one by one in the multivariate model. The multivariate model adjusted for age, parity, body mass index, fasting insulin, homeostatic model assessment for insulin resistance, quantitative insulin sensitivity check index.

[ii] ^bIndicates statistical significance (P<0.05).

Multivariate linear regression analysis

Among the variables investigated in the multivariate linear regression analysis, only diastolic blood pressure (β=0.64, 95% CI, 0.30 to 0.98; P<0.001) was found to be significantly associated with the levels of systolic blood pressure (Table II). Multivariate linear regression analysis for diastolic blood pressure revealed that fasting insulin levels (β=0.48; 95% CI, 0.02 to 0.94; P=0.040), HOMA-IR (β=2.51; 95% CI, 0.08 to 0.4.95; P=0.043) were associated with the levels of diastolic blood pressure (Table III).

Discussion

The main finding of the present study was that the multivariate regression model revealed that fasting insulin levels, HOMA-IR and systolic blood pressure were significantly associated with the levels of diastolic blood pressure among normotensive and euglycemic pregnant women. The ability of HOMA-IR to predict the levels of diastolic blood pressure in the present study was similar to the results reported in the study by Furugen et al (16), who tried to predict the development of hypertension in a cohort of normotensive adult Japanese subjects. They found that HOMA-IR and the Matsuda-DeFronzo insulin sensitivity index (ISI-M) were associated with blood pressure. However, the ISI-M was found to be more sensitive than HOMA-IR in the prediction of hypertension, which indicates an extra-hepatic source of IR (16). However, the present study did not calculate the ISI-M of the participants; otherwise, the findings could differ. Perhaps the presence of the placenta in pregnant women represents an additional source of IR (17). This is supported by the findings reported in the study by Jacober et al (18), who investigated insulin indices with blood pressure among nulliparous preeclamptic women at 3 to 6 months postpartum and found no association. This may point to the transient effect of IR associated with the presence of the placenta during pregnancy. It is worth mentioning that it is considered that the placenta synthesizes adipokines, such as leptin and others, and mediates IR by affecting insulin hormone signaling and initiating cascades of events in endothelial cells, leading to preeclampsia (19).

In the present study, HOMA-IR and fasting insulin levels were found to be associated with blood pressure. This is consistent with a previous study by Kazumi et al (20), which compared insulin-resistant indices in young normotensive subjects and subjects who were not hypertensive, but had high blood pressure readings. Hyperinsulinemia associated with IR (1) may indicate that the secretory function of β-cells remains responsive as long as the blood pressure readings remain within the normal range, as in the present study sample. However, this sustained hyperinsulinemia state may precipitate elevated blood pressure. A number of hypotheses have been postulated to explain the mechanism of hyperinsulinemia in the pathogenicity of hypertension. It has been mentioned that hyperinsulinemia leads to the dysfunction of immune cells, particularly T-cells and macrophages, a cardinal pathological mechanism in preeclampsia (21). Through this mechanism, the activity of Th1/Th2 cells is increased in the maternal circulation and the fetus-placenta interface (21). In addition to modulating the immune system, hyperinsulinemia triggers the sympathetic tone of the nervous system, increases renal sodium retention and imbalances the cation distribution across the membrane (6,7). Collectively, these mechanisms at least partially contribute to the pathogenicity of hypertension. However, it is uncertain why the patients in the present study did not develop hypertension or preeclampsia. Perhaps they did not reach the threshold at which the endothelial changes become irreversible. Long-standing and pre-gestational IR is considered a risk factor for preeclampsia (22). However, the exact turning point from normotension to hypertension is uncertain. This issue remains unresolved, whether during pregnancy or not.

In conclusion, measuring IR indices is not a routine antenatal care test. However, the current findings shed light on the association between HOMA-IR, fasting insulin levels and blood pressure readings. Although the blood pressure readings of the participants in the present study were still within the normal range, strict follow-up, close monitoring and timely intervention may alleviate major maternal-fetal cardiovascular complications. The present study was a single center initial study, and had certain limitations that need be addressed for a clear interpretation of our findings. Firstly, the design was a cross-sectional design; therefore, causality cannot be inferred; in addition, due to design, the authors could not determine whether the onset of IR in the study sample was before pregnancy or during pregnancy. Additionally, gestational complications, such as preeclampsia and gestational diabetes, were not addressed. Secondly, some confounding factors, such as a family history of diabetes/hypertension were not enquired. Therefore, further multi-center studies with a longitudinal design are required to measure more IR indices, such as ISI-M to establish/exclude extra-hepatic IR and detect pregnancy related complications, such as preeclampsia and gestational diabetes.

Acknowledgements

Not applicable.

Funding

Funding: No funding was received.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

ETEE and DAR were involved in the conceptualization of the study. IA, HZH and ASA were involved in the study methodology and design. ASA, DAR and IA were involved in data curation. ETEE, HZH and IA were involved in the formal analysis. ASA and HZH were involved in the investigative and procedural aspects of the study. DAR, IA and ETEE confirm the authenticity of all the raw data. ASA, DAR, HZH, ETEE and IA were involved in the drafting and reviewing of the primary version of the manuscript. All authors have read and approved the final manuscript.

Ethics approval and consent to participate

The study protocol was approved by the Research Ethics Board of the Department of Obstetrics and Gynecology, Faculty of Medicine, University of Khartoum and accordingly Ethical Approval was issued under number (#2020, 08). Informed written consent was obtained after explaining the research objectives. All experiments, sample collection and handling of patient data were conducted in accordance with the Declaration of Helsinki.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References