Abstract
Purpose:
Stress and elevated maternal glycemia have negative effects on pregnancy. We evaluated the association of hair cortisol concentrations (HCC), a marker of chronic stress, with insulin resistance and gestational diabetes (GDM).
Methods:
527 women from Lima, Peru, provided a hair sample in the second trimester of their pregnancy to measure HCC using liquid chromatography-tandem mass spectrometry. Each 6 cm of hair captured HCC in early (T1=1-12 weeks) and mid-pregnancy (T2=13-24 weeks). Routine screening was implemented to diagnose GDM in mid-pregnancy (IADPSG criteria). Multivariable linear and logistic regression models mutually adjusted for putative risk factors, including maternal sociodemographic factors, diabetes history, and hair characteristics, were used to estimate the association of HCC with GDM and other glycemic traits.
Results:
GDM was diagnosed in 122 (23%) women. Mean HCC across pregnancy was T1=3.7 (± 3.4) pg/mg and T2=4.8 (± 3.4) pg/mg. HCC was associated with increased log-transformed units of fasting insulin [T1=0.15 (0.03,0.27), T2=0.17 (0.04,0.30)], HOMA-IR [T1=0.14 (0.01,0.26), T2=0.17 (0.03,0.30)], and HOMA-B [T1=0.20 (0.05,0.34), T2=0.20 (0.04,0.36)], but not with the odds of GDM [T1=0.95 (0.63,1.40), T2=1.11 (0.74,1.67)].
Conclusions:
Elevated maternal HCC was associated with abnormal insulin homeostasis in pregnancy. Dysregulation of the hypothalamic-pituitary-adrenal axis, as reflected by high HCC, may also contribute to the insulin resistance syndrome in pregnancy.
Keywords: Hair cortisol, GDM, fasting insulin, HOMA scores, pregnancy
1. Introduction
Pregnancy is one of the most sensitive periods of development for the offspring (1). According to fetal programming theories, intrauterine conditions during pregnancy play a major role in shaping offspring health across the lifespan and intergenerationally (2). Chronic stress, prolonged or excessive stress, during pregnancy, is associated with detrimental health consequences (3) for the mother (4), and for the offspring (5), including the risk of neurodevelopmental problems, lower birth weight, shorter gestational age at delivery, and risk of cardiometabolic diseases (6). One of the well-established mechanisms by which chronic stress affects health is through the dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, which is often measured by the levels of cortisol (7).
Cortisol, the primary end-product of the HPA axis response to stress, is a glucocorticoid secreted by the adrenal glands. It is subject to variations in interindividual fluctuations, like circadian rhythm, mood changes, momentary events, or study procedures (8). Situational factors that affect cortisol concentrations like alcohol (9) and nicotine use (10), physical activity (11), food intake (12), injury (13), and time of the day (14), are known to affect cortisol sampled in blood, saliva, and urine, but not in hair (8). Cortisol concentrations measured in hair offer a non-invasive, longer retrospective assessment (weeks to months) of cortisol synthesis and secretion (8) compared to cortisol measured in other body fluids, and it has been established as a suitable biomarker of chronic stress and HPA axis dysregulation (7). For instance, cortisol measured in the saliva follows a diurnal pattern, whereby measures taken at awakening are 20%-50% higher than those taken post-awakening when they tend to decrease (15). Because of this short-term representation of cortisol when using saliva, blood, and urine samples, repeated measures are required to assess long-term cortisol levels, but this is more restrictive in epidemiological studies. Thus, hair cortisol has become the biomarker of choice to assess chronically elevated levels of cortisol in relation to adverse health outcomes (8).
Despite the well-known association of cortisol with incidence (16) and prevalence (17) of diabetes, and with glycemic traits in non-pregnant populations (18), less is known about the association of cortisol with gestational diabetes mellitus (GDM). Knowing the impact that GDM has on adverse maternal and infant health outcomes (19), it is important to understand how dysregulation of the HPA-axis, reflected by elevated cortisol concentrations during specific periods in pregnancy, may be associated with GDM and with related glycemic measures (20). We designed the present study to understand the extent to which maternal hair cortisol concentrations in early and mid-pregnancy are associated with measures of glucose tolerance in pregnancy and potentially with GDM using data from a well-characterized population in Peru. Our study objectives were to examine the relation between hair cortisol concentrations in early and mid-pregnancy with i) GDM status; and ii) glucose traits, insulin, and indices of insulin resistance and β-cell function. We hypothesized that elevated hair cortisol concentrations in early and mid-pregnancy, indicative of HPA axis dysregulation, are positively associated with increased insulin resistance and odds of GDM.
2. Methods
2.1. Study population
We analyzed hair cortisol data from a sub-study of the Screening Treatment and Effective Management of Gestational Diabetes Mellitus (STEM-GDM) study. The STEM-GDM is a cohort study designed to investigate the prevalence of GDM and provide evidence to improve local guidelines for the screening, management and treatment of GDM among pregnant Peruvian women attending the Instituto Nacional Materno Perinatal (INMP) in Lima, Peru (21). The initial sample was 1,300 participants recruited between February 2013 and June 2014. Study details have been previously described (21). Briefly, eligible women were those who initiated prenatal care prior to 28 weeks of gestation, had completed an oral glucose tolerance test (OGTT), and had singletons. Women were excluded if they were younger than 18 years of age, were not able to speak, read and write in Spanish, or were not planning to deliver at INMP. They were also excluded if they had previous records of existing diabetes, use of medication to treat glucose intolerance, or were receiving treatment for any other chronic condition (21). In total, 527 women provided hair samples and were included in this study. Participants included in the study were not different from the overall sample in most of their demographic and lifestyle characteristics, except for values of age, pre-pregnancy BMI, and family history of diabetes, which were higher in women included compared to those excluded from the study due to lack of hair samples (Supplementary Table A.1). All study participants provided written informed consent. The Institutional Review Board of the INMP, Lima, Peru, and the Harvard T.H. Chan School of Public Health Office of Human Research Administration, Boston, MA, USA, approved the study protocols and procedures.
2.2. Sociodemographic measures and covariates
We collected maternal sociodemographic and lifestyle factors, medical and reproductive history using in-person structured interviews between 24 and 28 weeks of pregnancy. The sociodemographic variables included maternal age, ethnicity (mestizo vs. no-mestizo), marital status (married or living with a partner vs. single or divorced), paid employment (yes vs. no), maternal level of education (≤6, 7-12, >12 years), food insecurity (yes vs. no), family history of diabetes (FHD) among first-degree relatives, and perceived health during pregnancy (good vs. poor). We also have information about maternal parity (primiparous vs. multiparous), gestational week at enrollment, alcohol use, and smoking use during pregnancy (yes vs. no). Pre-pregnancy body mass index (BMI) was calculated dividing weight (kilograms) by the square of height (m2). We classified mothers into three BMI groups: normal weight (< 25 kg/m2), overweight (25-29.9 kg/m2), and obese (≥30 kg/m2). There were seven underweight women (<18.5 kg/m2) included in the same category as those with normal weight as excluding them did not alter the analysis. Hair characteristics were also collected, including hair color (black, brown, other), hair structure (curly vs. straight), hair washing frequency, use of shampoo and/or conditioner, hair chemical treatment use, and hair cutting frequency.
2.3. Glucose and Insulin measures in the OGTT
GDM diagnosis was determined using a standardized oral glucose tolerance test (OGTT) administered in mid-pregnancy (mean 25.6 weeks ± 1.3 standard deviations) and following IADPSG criteria (22). During the OGTT, venous blood was drawn 8 hours after an overnight fasting, and 1-h and 2-h after receiving a sugary drink with 75-g of anhydrous oral glucose dissolved in water. Plasma glucose (mg/dl) was measured at fasting, 1-h and 2-h post glucose load using the glucose oxidase-peroxidase method in duplicate with an auto-analyzer for biochemical tests (21). Fasting plasma insulin (μIU/mL) was measured using an electrochemical luminescence immunoassay (Roche Diagnostics, Mannheim, Germany) in 328 (62.2%) of the 527 women with complete OGTT and HCC data. We compared baseline characteristics between women with and without available fasting insulin measures. Characteristics were overall comparable between both study populations, except for a higher proportion of women with black hair and with a female child in the group with measured fasting insulin.
We calculated indices of insulin homeostasis in pregnancy using samples with complete fasting insulin and fasting glucose data (n=328). We calculated the homeostasis model assessment for insulin resistance (HOMA-IR) by dividing the product of fasting plasma glucose (mg/dl) and fasting plasma insulin (μIU/mL) by 405 (23). HOMA-IR values are inversely proportional to insulin sensitivity, and higher HOMA-IR values indicate higher insulin resistance (24). We calculated HOMA for β-cell function (HOMA-B) using the following formula (fasting plasma insulin x 20) / [(fasting plasma glucose/18) - 3.5] (23). Higher values of the HOMA-B score indicate better β-cell function.
2.4. Hair sample collection and cortisol analysis
The procedures used to collect hair samples and extract hair cortisol concentration (HCC) have been described earlier (25). Briefly, trained research staff collected hair samples from the posterior vertex region (top of the head), as close to the scalp as possible, at the time of enrollment in mid-pregnancy. Cortisol measured in the posterior vertex shows less variability over time in cortisol levels compared to other areas of the scalp (15). Before laboratory analysis, each hair sample was segmented into two 3 cm hair segments. A 3-cm hair sample is commonly used in the literature to reflect the retrospective activity of the HPA axis over the past 3 months (15, 26, 27). Considering an average human hair growth of 1 cm per month (28), the hair segment taken 3-6 cm from the scalp (HCC segment II) reflected maternal cortisol concentrations in the first trimester (1-12 weeks), and the one taken 3 cm from the scalp (HCC segment I) reflected cortisol concentrations in mid-pregnancy (13-24 weeks). Hair segments from the same woman were analyzed in the same batch to reduce technical variability, and variability within batches was determined using six randomly allocated quality control samples (25). The segmented hair samples were washed in isopropanol twice for three minutes, dried, weighed out to 7.5 mg of whole hair cut into small pieces, and incubated in methanol to extract cortisol (pg/mg) using a standardized liquid chromatography-tandem mass spectrometry assay, with a lower limit of detection of 0.1 mg/pg (29). One sample showing extreme values of HCC across pregnancy (HCC > 40 pg/mg) was excluded from the study. Thus, complete cortisol values were present for 527 samples in segment I (mid-pregnancy) and 525 in segment II (early-pregnancy) for further analyses. All laboratory assays were completed without knowledge of the participants’ medical history.
2.5. Statistical analysis
We inspected the distribution of exposures and outcomes using the Shapiro normality test, histograms, and QQ-plots. Samples with missing values for covariates of interest like pre-pregnancy BMI (n=13), gestational week (n=2), and hair tinting (n=43), were further excluded from the analysis. Since the distribution of HCC, fasting glucose, fasting insulin, and the HOMA scores was skewed, we used the natural log-transformation of their values to fulfill model assumptions. We described sociodemographic characteristics of study participants using means and standard deviations (SD) or the number and percent (%). We provided means and SD of log-transformed traits in their original units for descriptive purposes. We used the Wilcoxon rank sum test, the Fisher's exact test or the Pearson's Chi-squared test, to compare characteristics between GDM groups. We analyzed the distribution of HCC across categories of maternal characteristics using the mean and SD and reported P-values of significance using the student's t-test and the ANOVA test.
We implemented unadjusted and adjusted regression models with log-transformed values of HCC as the independent variable, to estimate the associations with glycemic traits (i.e., fasting glucose, glucose-1h, glucose-2h and fasting insulin), HOMA scores and GDM status. In a basic model, associations were adjusted for maternal age, parity, and pre-pregnancy BMI. In fully adjusted models, we additionally adjusted for gestational week, hair tinting, FHD, maternal education and food insecurity. HCC was analyzed continuous and by tertiles. To ease interpretation of results, we showed effect estimates of normalized outcomes in their original units. Effect estimates were interpreted as the average change in the trait, or the log-odds ratio of GDM, per unit increase in log-transformed values of HCC. We presented effect estimates and 95% confidence intervals (95% CI) in tables. All analyses were conducted in R version 4.1.1, and tests were two-tailed and deemed statistically significant at P < 0.05.
2.6. Sensitivity analyses
We performed an inverse probability weighting analysis to account for potential selection due to differences observed in relevant covariates between participants included and those excluded from the HCC study. This analysis balances the distribution of measured confounders across levels of the treatment variable (HCC in early and mid-pregnancy) using weights calculated in the total sample. We rerun analyses for the fully adjusted model, including weights, and we compared these results with the primary analysis (without weights). Evidence of selection bias was considered if the correlation between estimates of the two analyses was small (r < 0.3, P < 0.05). Our primary analysis results are based on a complete case analysis. However, to account for potential bias in the estimates due to missing values for certain covariates (up to 8% missingness), we rerun regressions using a dataset derived from a multiple imputation analysis using the mice (v.3.15.0) R package (30).
3. Results
3.1. Characteristics of study participants
The baseline characteristics of study participants are presented in Table 1. The mean maternal age was 29 (± 6.2) years, and mean gestational age at HCC measurement was 25 (± 1.3) weeks. Most of the study participants were mestizo (99%), multiparous (69%), married or lived with their partners (86%), and reported food insecurity (68%).
Table 1.
Maternal characteristics overall and by GDM status among pregnant women in Lima, Peru (N=527).
| Characteristics | All women (N= 527) n(%) |
Controls (N=405) n(%) |
GDM cases (N=122) n(%) |
P |
|---|---|---|---|---|
| Maternal age (years) a | 29.5 (6.2) | 29.2 (6.0) | 30.4 (6.7) | 0.077 |
| Maternal ethnicity | ||||
| Mestizo | 522 (99) | 400 (99) | 121 (99) | > 0.9 |
| No-Mestizo | 6 (1.1) | 5 (1.3) | 1 (0.8) | |
| Gestational week a | 25.6 (1.3) | 25.5 (1.3) | 25.7 (1.3) | 0.2 |
| Pre-pregnancy BMI (kg/m2) a | 25.9 (4.2) | 25.9 (4.2) | 25.7 (4.2) | 0.3 |
| Parity | ||||
| Primiparous | 162 (31) | 129 (32) | 33 (27) | 0.3 |
| Multiparous | 365 (69) | 276 (68) | 89 (73) | |
| Marital status | ||||
| Married/living with partner | 453 (86) | 348 (86) | 105 (86) | > 0.9 |
| Single/divorced | 74 (14) | 57 (14) | 17 (14) | |
| Employment, yes % | 181 (34) | 144 (36) | 37 (30) | 0.3 |
| Education (years) | ||||
| ≤6 | 12 (2.3) | 10 (2.5) | 2.0 (1.6) | > 0.9 |
| 7-12 | 279 (53) | 215 (53) | 64 (52) | |
| >12 | 236 (45) | 180 (44) | 56 (46) | |
| Smoking in pregnancy, yes % | 2.0 (0.4) | 1.0 (0.2) | 1 (0.8) | 0.4 |
| Alcohol use in pregnancy, yes % | 11 (2.1) | 8.0 (2.0) | 3.0 (2.5) | 0.7 |
| HCC I (mid-pregnancy) (pg/mg) | 4.8 (3.4) | 4.8 (3.5) | 5.0 (3.4) | 0.6 |
| HCC II (early-pregnancy) (pg/mg) | 3.7 (3.5) | 3.7 (3.5) | 3.9 (3.4) | 0.2 |
| Hair color | ||||
| Black | 396 (82) | 339 (84) | 57 (72) | 0.003 |
| Brown | 86 (18) | 66 (16) | 20 (25) | |
| Other | 2.0 (0.4) | 0 (0) | 2 (2.5) | |
| Natural hair structure | ||||
| Curly | 25 (5.2) | 20 (4.9) | 5.0 (6.3) | 0.6 |
| Straight | 459 (95) | 385 (95) | 74 (94) | |
| Hair tinting, yes % | 194 (40) | 158 (39) | 36 (46) | 0.3 |
| Hair dyeing, yes % | 5.0 (1) | 5 (1.2) | 0 (0) | > 0.9 |
| Food insecurity, yes % | 360 (68) | 276 (68) | 84 (69) | 0.9 |
| Perceived health in pregnancy | ||||
| Good | 269 (51) | 209 (52) | 60 (49) | 0.6 |
| Poor | 258 (49) | 196 (48) | 62 (51) | |
| Child sex | ||||
| Female | 215 (44) | 156 (42) | 59 (50) | 0.09 |
| Male | 277 (56) | 219 (58) | 59 (50) | |
| Family history of Diabetes, yes % | 210 (42) | 157 (41) | 53 (45) | 0.4 |
| Fasting plasma glucose (mg/dL) a | 84.0 (9.3) | 80.5 (6.7) | 95.5 (7.3) | <0.001 |
| Glucose-1h (mg/dL) a | 134.7 (22.9) | 129.5 (19.4) | 151.8 (25.2) | <0.001 |
| Glucose-2h (mg/dL) a | 103.6 (20.7) | 99.1 (17.0) | 118.7 (24.6) | <0.001 |
| Fasting insulin (μIU/mL) a | 5.3 (4.5) | 4.7 (4.0) | 6.7 (5.1) | <0.001 |
| HOMA-IR a | 1.1 (1.0) | 0.9 (0.8) | 1.6 (1.2) | <0.001 |
| HOMA-B a | 98.2 (92.5) | 107.8 (102.9) | 77.1 (59.1) | <0.001 |
Continuous variables were summarized using mean (SD). BMI: body mass index, HOMA-IR: homeostasis model assessment for insulin resistance, HOMA-B: homeostasis model assessment for β-cell function. P-value comparing maternal characteristics by GDM status was calculated using a Wilcoxon rank sum test for continuous non-parametric variables, and the Fisher's exact test or Pearson's Chi-squared test for categorical variables.
A total of 122 (23%) participants received a GDM diagnosis using an OGTT-administered test, and they were referred to receive standard care for glucose management and monitoring. The mean HCC in early (segment II) and mid-pregnancy (segment I) was 3.7 (± 3.5) mg/pg and 4.8 (± 3.4) mg/pg, respectively (Supplementary Fig. A.1), and both measures showed a strong positive correlation (r= 0.67, P < 0.01) (Figure 1). Participants with and without GDM were similar in sociodemographic, lifestyle factors, and hair characteristics, except for hair color (more variety of hair colors in GDM group). As expected, participants with GDM were more likely than those without to have higher mean levels of OGTT glucose values (fasting, glucose-1h and glucose-2h), higher fasting insulin and HOMA-IR, but lower HOMA-B values (P < 0.001) (Table 1).
Figure 1.
Correlogram showing the unadjusted associations between maternal hair cortisol concentrations from two hair segments, and glucose and insulin traits assessed in mid-pregnancy during a 75-g OGTT. The legend color represents the direction and strength of the pairwise correlations. P-value was > 0.05 in the pairwise correlations between HCC (two segments) and fasting glucose, glucose-1h and glucose-2h. FG norm: normalized fasting glucose (mg/dL), FI_norm: normalized fasting insulin (uIU/mL), HOMA-IR norm: normalized homeostasis model assessment for insulin resistance, HOMA-B_norm: normalized homeostasis model assessment for β-cell function, HCC-I norm: normalized hair cortisol concentrations in mid-pregnancy, HCC-II_norm: normalized hair cortisol concentrations in early-pregnancy.
3.2. Levels of HCC differed across categories of BMI and insulin traits in pregnancy
The distribution of HCC in pregnancy across maternal characteristics is shown in Table 2. For both pregnancy periods (early and mid-pregnancy), we observed that levels of HCC were moderately higher (> 16%) in women with overweight or obesity status compared to those with normal weight (P < 0.001). Overall, HCCs tended to increase with increasing values of fasting insulin, HOMA-IR, and HOMA-B (P ≤ 0.03). Levels of HCC in early pregnancy (1-12 weeks) were significantly higher in women of older age (P=0.01) and in those who reported alcohol consumption in pregnancy (P=0.03). We did not observe statistically significant differences in HCC by any other covariate.
Table 2.
Hair cortisol concentrations (HCC) in early and mid-pregnancy across characteristics of pregnant women in Lima, Peru (N=527).
| Characteristics | N | HCC early-pregnancy | P | N | HCC mid-pregnancy | P |
|---|---|---|---|---|---|---|
| Mean (SD) | Mean (SD) | |||||
| Sociodemographic traits | ||||||
| Maternal age (years) | ||||||
| ≤ 19 | 13 | 3.24(2.32) | 0.01 | 13 | 5.39(3.54) | 0.12 |
| 20-29 | 251 | 3.32(3.11) | 253 | 4.59(3.26) | ||
| 30-34 | 132 | 4.25(3.69) | 132 | 5.28(3.47) | ||
| ≥ 35 | 129 | 4.02(3.85) | 129 | 4.86(3.72) | ||
| Gestational week | ||||||
| < 25 weeks | 97 | 3.56(2.64) | 0.38 | 97 | 5.00(3.33) | 0.27 |
| 25-26 weeks | 124 | 3.62(3.52) | 125 | 4.67(3.33) | ||
| ≥ 27 weeks | 302 | 4.03(3.82) | 303 | 5.17(3.79) | ||
| Pre-pregnancy BMI | ||||||
| < 25 (normal weight) | 244 | 3.34(3.13) | <0.001 | 244 | 4.41(3.08) | <0.001 |
| 25 – 29.9 (overweight) | 75 | 3.90(3.58) | 75 | 5.13(3.85) | ||
| ≥ 30 (obese) | 188 | 4.89(4.15) | 190 | 5.86(3.45) | ||
| Smoking during pregnancy | ||||||
| No | 523 | 3.72(3.46) | 0.79 | 525 | 4.84(3.44) | 0.59 |
| Yes | 2 | 4.45(3.6) | 2 | 6.38(4.04) | ||
| Alcohol during pregnancy | ||||||
| No | 514 | 3.70(3.47) | 0.03 | 516 | 4.81(3.44) | 0.06 |
| Yes | 11 | 4.57(2.18) | 11 | 6.64(3.31) | ||
| Perceived Health in pregnancy | ||||||
| Good | 268 | 3.65(3.33) | 0.35 | 269 | 4.79(3.53) | 0.36 |
| Poor | 257 | 3.80(3.58) | 258 | 4.91(3.35) | ||
| Food insecurity | ||||||
| No | 166 | 3.48(2.92) | 0.49 | 167 | 4.69(3.22) | 0.88 |
| Yes | 359 | 3.83(3.67) | 360 | 4.92(3.54) | ||
| Family history of diabetes | ||||||
| No | 22 | 3.72(3.58) | 0.85 | 22 | 4.82(3.64) | 0.65 |
| Yes | 294 | 3.73(3.32) | 295 | 4.88(3.16) | ||
| Don’t know | 209 | 3.66(3.04) | 210 | 4.89(3.37) | ||
| Child sex | ||||||
| Female | 215 | 3.54(2.69) | 0.45 | 215 | 4.94(3.65) | 0.86 |
| Male | 275 | 3.85(4.04) | 277 | 4.82(3.36) | ||
| Hair characteristics | ||||||
| Hair color | ||||||
| Black | 395 | 3.43(2.87) | 0.16 | 396 | 4.74(3.35) | 0.66 |
| Brown | 86 | 4.57(4.87) | 86 | 5.19(4.03) | ||
| Other | 2 | 2.77(0.50) | 2 | 5.47(4.51) | ||
| Natural hair structure | ||||||
| Curly | 25 | 2.76(2.01) | 0.06 | 25 | 3.86(2.53) | 0.1 |
| Straight | 458 | 3.68(3.39) | 459 | 4.87(3.52) | ||
| Hair tinting | ||||||
| No | 289 | 3.66(3.37) | 0.86 | 290 | 4.79(3.26) | 0.9 |
| Yes | 194 | 3.58(3.29) | 194 | 4.87(3.79) | ||
| Hair dyeing | ||||||
| No | 478 | 3.62(3.34) | 0.31 | 479 | 4.81(3.47) | 0.74 |
| Yes | 5 | 4.57(2.38) | 5 | 5.86(4.44) | ||
| GDM and glycemic traits | ||||||
| Gestational diabetes | ||||||
| GDM | 121 | 3.89(3.43) | 0.34 | 122 | 5.02(3.42) | 0.44 |
| Controls | 404 | 3.67(3.46) | 405 | 4.80(3.45) | ||
| Fasting glucose (mg/dL) | ||||||
| < 92 mg/dl | 413 | 3.68(3.48) | 0.34 | 414 | 4.83(3.53) | 0.64 |
| ≥ 92 mg/dl | 112 | 3.86(3.36) | 113 | 4.90(3.11) | ||
| Glucose-1h (mg/dL) | ||||||
| < 180 mg/dl | 507 | 3.72(3.46) | 0.72 | 509 | 4.81(3.39) | 0.2 |
| ≥180 mg/dl | 18 | 3.87(3.35) | 18 | 5.87(4.63) | ||
| Glucose-2h (mg/dL) | ||||||
| < 153 mg/dl | 514 | 3.71(3.44) | 0.48 | 516 | 4.82(3.38) | 0.27 |
| ≥ 153 mg/dl | 11 | 4.42(3.96) | 11 | 6.37(5.59) | ||
| Fasting insulin (μIU/mL) | ||||||
| 0.26-3.27 | 110 | 2.96(3.10) | 0.01 | 110 | 3.89(3.31) | 0.01 |
| 3.29-5.28 | 108 | 3.50(3.90) | 109 | 4.21(2.72) | ||
| 5.31-45.57 | 108 | 3.80(3.41) | 109 | 4.70(2.90) | ||
| HOMA-IR | ||||||
| 0.05-0.66 | 110 | 2.87(2.86) | 0.01 | 110 | 3.75(2.82) | 0.02 |
| 0.67-1.13 | 108 | 3.69(4.10) | 109 | 4.47(3.22) | ||
| 1.14-9.11 | 108 | 3.70(3.37) | 109 | 4.58(2.89) | ||
| HOMA-B | ||||||
| 5.59-52.86 | 109 | 2.69(1.65) | 0.03 | 109 | 3.55(2.09) | 0.01 |
| 53.43-102.81 | 109 | 3.82(4.59) | 110 | 4.76(3.71) | ||
| 103.08-911.32 | 108 | 3.73(3.50) | 109 | 4.49(2.85) |
Complete data for HCC in early and mid-pregnancy was available in N=525 and 527 women, respectively. Total sample may be lower for some variables due to missingness (gestational week n=2, BMI n=13, child sex=35, hair characteristics n=43). Fasting insulin and HOMA scores were available for a subset of 328 women with HCC data. BMI: body mass index, HOMA-IR: homeostasis model assessment for insulin resistance, HOMA-B: homeostasis model assessment for β-cell function, SD: standard deviation. P-value difference in the mean of log-transformed HCC values in pregnancy across maternal characteristics was calculated using a student’s t-test (binary variables) or an ANOVA test (categorical ordinal variables).
3.3. Elevated HCC was positively associated with insulin resistance and measures of beta-cell function
Using Pearson's correlations, we observed significant positive associations of HCC in early and mid-pregnancy with fasting insulin, HOMA-IR, and HOMA-B scores (P range 0.001-0.01), but the magnitude of these correlations was relatively small (r < 0.2) (Figure 1, Supplementary Fig. A.2-A.4).
We observed statistically significant associations between HCC and maternal fasting insulin and HOMA scores across pregnancy. After adjusting for potential confounders, a one-unit increase in log-transformed values of HCC in early pregnancy was associated with higher log-transformed units of fasting insulin (beta=0.15, 95% CI = 0.03, 0.27, P = 0.02), HOMA-IR (a measure of insulin resistance) (beta=0.14, 95% CI=0.01, 0.26, P = 0.03), and HOMA-B (a measure of β-cell function) (beta=0.20, 95% CI= 0.05, 0.34, P = 0.01) (Table 3). Similarly, a one-unit increase in log-transformed values of HCC in mid-pregnancy was associated with an increase in log-transformed units of fasting insulin (beta=0.17, 95% CI=0.04, 0.30, P = 0.01), HOMA-IR (beta=0.17, 95% CI=0.03, 0.30, P = 0.02), and HOMA-B (beta=0.20, 95% CI=0.04, 0.36, P = 0.01) (Table 3). When HCC was analyzed by tertiles, we observed a similar trend towards higher values of insulin and the HOMA scores in the upper versus the lower tertile (reference level) of HCC assessed in mid-pregnancy (P ≤ 0.01, Appendix Table A.2); in early pregnancy, the opposite trend was observed in associations with HCC.
Table 3.
Associations of maternal hair cortisol concentrations in early and mid-pregnancy, with glycemic traits, HOMA scores and odds of GDM among pregnant women in Lima, Peru (N=527).
| HCC early-pregnancy (1-12 gestational weeks) | HCC mid-pregnancy (13-24 gestational weeks) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| N | Beta (95% CI) |
Beta (95% CI) | P* | Adjusted R- square |
N | Beta (95% CI) | Beta (95% CI) | P* | Adjusted R- square |
|
| Unadjusted model | ||||||||||
| Fasting Insulin | 326 | 0.19 (0.07, 0.30) | 0.99 (0.26, 1.73) | 0.002 | 0.028 | 328 | 0.22 (0.09, 0.35) | 1.05 (0.22, 1.87) | 0.001 | 0.029 |
| HOMA-IR | 326 | 0.20 (0.08, 0.31) | 0.23 (0.07, 0.40) | 0.002 | 0.028 | 328 | 0.22 (0.09, 0.36) | 0.25 (0.06, 0.43) | 0.001 | 0.029 |
| HOMA-B | 326 | 0.16 (0.03, 0.29) | 13.78 (−1.51, 29.07) | 0.013 | 0.016 | 328 | 0.19 (0.05, 0.33) | 12.01 (−5.19, 29.21) | 0.010 | 0.017 |
| Fasting glucose | 525 | 0.001(−0.01, 0.02) | 0.08 (−1.16, 1.32) | 0.859 | −0.002 | 527 | 0.002 (−0.01, 0.02) | 0.07 (−1.25, 1.38) | 0.829 | −0.002 |
| Glucose-1h | 525 | 2.17 (−0.87, 5.20) | --- | 0.162 | 0.002 | 527 | 2.93 (−0.29, 6.15) | --- | 0.075 | 0.004 |
| Glucose-2h | 525 | 2.65 (−0.10, 5.39) | --- | 0.060 | 0.005 | 527 | 2.60 (−0.32, 5.53) | --- | 0.081 | 0.004 |
| GDM† | 525 | 0.15 (−0.16, 0.46) | --- | 0.331 | --- | 527 | 0.13 (−0.20, 0.46) | --- | 0.44 | --- |
| Basic model | ||||||||||
| Fasting Insulin | 319 | 0.17 (0.06, 0.28) | 0.89 (0.16, 1.62) | 0.003 | 0.103 | 321 | 0.19 (0.07, 0.32) | 0.93 (0.11, 1.74) | 0.003 | 0.103 |
| HOMA-IR | 319 | 0.18 (0.06, 0.29) | 0.21 (0.05, 0.37) | 0.003 | 0.103 | 321 | 0.20 (0.07, 0.33) | 0.22 (0.04, 0.40) | 0.003 | 0.104 |
| HOMA-B | 319 | 0.15 (0.02, 0.28) | 12.80 (−2.62, 28.23) | 0.020 | 0.049 | 321 | 0.17 (0.03, 0.31) | 10.25 (−7.06, 27.55) | 0.021 | 0.049 |
| Fasting glucose | 512 | 0.00 (−0.01, 0.01) | −0.02 (−1.29, 1.25) | 0.997 | 0.004 | 514 | 0.002 (−0.01, 0.02) | 0.12 (−1.22, 1.46) | 0.788 | 0.004 |
| Glucose-1h | 51 | 1.58 (−1.52, 4.68) | --- | 0.317 | 0.017 | 514 | 2.75 (−0.51, 6.02) | --- | 0.099 | 0.021 |
| Glucose-2h | 512 | 2.19 (−0.55, 4.92) | --- | 0.118 | 0.032 | 514 | 2.74 (−0.15, 5.62) | --- | 0.064 | 0.034 |
| GDM† | 512 | 0.17 (−0.15, 0.48) | --- | 0.284 | --- | 514 | 0.19 (−0.15, 0.53) | --- | 0.278 | --- |
| Fully adjusted model | ||||||||||
| Fasting Insulin | 278 | 0.15 (0.03, 0.27) | 0.69 (−0.03, 1.41) | 0.016 | 0.134 | 279 | 0.17 (0.04, 0.30) | 0.69 (−0.09, 1.47) | 0.010 | 0.137 |
| HOMA-IR | 278 | 0.14 (0.01, 0.26) | 0.14 (−0.01, 0.30) | 0.031 | 0.133 | 279 | 0.17 (0.03, 0.30) | 0.15 (−0.02, 0.32) | 0.015 | 0.137 |
| HOMA-B | 278 | 0.20 (0.05, 0.34) | 17.20 (−0.62, 35.01) | 0.008 | 0.063 | 279 | 0.20 (0.04, 0.36) | 11.99 (−7.45, 31.45) | 0.012 | 0.059 |
| Fasting glucose | 469 | −0.01 (−0.02, 0.01) | −0.78 (−2.10, 0.53) | 0.259 | 0.008 | 470 | −0.002 (−0.02, 0.01) | −0.24 (−1.59, 1.12) | 0.803 | 0.005 |
| Glucose-1h | 469 | 1.05 (−2.27, 4.37) | --- | 0.535 | 0.016 | 470 | 2.09 (−1.32, 5.50) | --- | 0.230 | 0.018 |
| Glucose-2h | 469 | 1.91 (−0.99, 4.81) | --- | 0.198 | 0.019 | 470 | 2.33 (−0.64, 5.31) | --- | 0.125 | 0.021 |
| GDM† | 469 | −0.05 (−0.46, 0.34) | --- | 0.808 | --- | 470 | 0.11 (−0.30, 0.51) | --- | 0.608 | --- |
GDM: gestational diabetes mellitus, HOMA-IR: homeostasis model assessment for insulin resistance, HOMA-B: homeostasis model assessment for β-cell function. Basic model: model adjusted for maternal age, parity, and pre-pregnancy BMI. Fully adjusted model: additionally adjusted for gestational week, hair tinting, family history of diabetes, maternal education and food insecurity. Beta: association estimates obtained using original units for outcomes that were log-transformed (i.e., fasting insulin (μIU/mL), HOMA scores and fasting glucose (mg/dL)).
P value obtained from the model on the logarithmic scale for log-transformed traits; otherwise, from the model using the trait in their original scale.
Estimates for GDM are shown in the log-odds scale. Missing values in our sample were seen for pre-pregnancy BMI (n=13), gestational week (n=2) and hair tinting (n=43).
In a sensitivity analysis using IPWs, the effect sizes were not materially different from those obtained in our primary analysis (Appendix Table A.3). Estimates of the fully adjusted model with and without weights showed a high correlation (rho=1.0, P= 0.004). These results suggest no evidence of selection bias in our study. Estimates of the complete case analysis were consistent with those obtained using a dataset derived from a multiple imputation analysis (rho=0.82, P= 0.034) (Appendix Table A.4)
4. Discussion
In this cross-sectional study of pregnant Peruvian women, we showed that mean levels of HCC increased from early to mid-pregnancy. In both time periods, cortisol levels were positively associated with mid-pregnancy measures of fasting insulin, insulin resistance, and β-cell function. We did not observe an association of HCCs with odds of GDM or with glucose traits in the OGTT. Dose-response analyses confirmed a linear trend in the association of HCC with increasing insulin levels and the HOMA scores.
Our results showing an increase in maternal HCC levels from the first (T1) to the second (T2) trimester are consistent with previous studies (31, 32). Nonetheless, mean values of HCC identified in each trimester were lower than those reported in previous studies in Europeans (32, 33), but similar to those observed in our prior study of Peruvian women (25). Differences in measures of HCC between studies may be related to the performance of the distinct biochemical methods used to assess HCC (LC-MS/MS in our case vs immunoassay methods for others).
In our study, we observed that higher cortisol concentrations in pregnancy were associated with increased fasting insulin and insulin resistance measured by HOMA-IR, with slightly larger effects seen for the associations with cortisol measured in mid-pregnancy versus early pregnancy. Our results are consistent with the literature (34-36), and with previous studies showing a positive correlation between serum cortisol and HOMA-IR among pregnant women with or without GDM (37). The observed associations are explained by the metabolic changes that occur during pregnancy. In late pregnancy, to secure the transference of nutrients to the fetus and adequate fetal development, there is an increase in insulin resistance coupled with large insulin secretion, increased lipolysis, and reduced glucose uptake in maternal peripheral tissues (34). Euglycemic levels in pregnancy are maintained by the interplay between lactogenic hormones and glucocorticoids. Lactogenic hormones, including prolactin and placental lactogens, stimulate β-cell function and insulin secretion, while glucocorticoids promote insulin resistance (34). Cortisol, a type of glucocorticoid, is known for reducing glucose-induced insulin release, causing alterations in β-cell proliferation and function in pregnancy (34, 38). Cortisol also promotes hepatic glucose production and decreases peripheral insulin sensitivity (37, 39). All the above conditions aggravate the state of hyperglycemia and may lead to glucose intolerance and GDM. Thus, in view of increasing cortisol levels in pregnancy, maternal insulin production increases to avoid insulin resistance and maintain glucose tolerance (34, 35).
Despite showing a clear positive association between cortisol and insulin resistance in pregnancy, we found no evidence that cortisol influenced glucose levels at any point during the OGTT, or the odds of GDM. Our results of the null associations of HCCs with GDM are consistent with some prior studies. For instance Feng et al. (37) using serum cortisol, found slightly higher cortisol in the GDM compared to the non-GDM group, but this difference was not statistically significant (P > 0.1). In contrast, Ahmed et al. (40) found that higher serum cortisol was associated with GDM and with impaired glucose tolerance (IGT) in a sample of 90 pregnant women (30 GDM, 30 IGT cases, and 30 controls). Compared to previous studies, ours had more power to identify a putative association between GDM, OGTT glucose, and cortisol, considering our larger sample size (N= 527 vs < 140 in others), and the fact that we measured cortisol in hair, which provides a longer-term assessment of cortisol relative to the more situational measure of serum cortisol. Even though cortisol could promote factors related to glucose intolerance and GDM, including increased hepatic glucose production, β-cell dysfunction, and lower insulin secretion (37), its association with glucose metabolism in pregnancy is yet to be fully investigated. We hypothesize that HPA-dysfunction, reflected by elevated cortisol levels, may influence glucose intolerance in pregnancy through insulin resistance and β-cell dysfunction but not directly through changes in glucose levels.
Contrary to what we expected, higher values of the HOMA-B score, which indicate better β-cell function, were seen with increasing levels of cortisol, and the linear trend in this association was more conspicuous with cortisol measured in mid-pregnancy rather than in early pregnancy. One of the explanations for this finding may be related to a compensatory mechanism prompted by lactogenic hormones in response to lower insulin sensitivity in mid-pregnancy, partly due to higher cortisol levels. In this compensatory mechanism, the effect of the lactogenic hormones prolactin and placental lactogen, triggers a rise in insulin secretion derived from increased β-cell mass (34, 35). Thus, lactogen hormones may be important counteractors of the effect of cortisol on insulin resistance, especially in late pregnancy, improving maternal glucose tolerance by enhancing β-cell mass and function. It is worth noting that the direct effect of cortisol on β-cell function remains elusive. Nonetheless, studies outside of pregnancy and animal studies support the concept that cortisol-induced insulin resistance in peripheral tissues (i.e., muscle, liver, adipose tissue), triggers β-cell hyperplasia and hyperinsulinemia as a compensatory mechanism to maintain euglycemia (41). Thus, measures of β-cell function like the HOMA-B score may increase temporarily in response to hypercortisolism. However, long-term exposure to cortisol, like under chronic stress conditions, can exceed the coping response of β-cells, leading to lower insulin secretion, glucose intolerance, and diabetes (41). Our study's observation may reflect this transitional period between β-cell hyperactivity due to prolonged cortisol exposure, and β-cell dysfunction occurring later in pregnancy. Therefore, having measures of HOMA-B throughout pregnancy, may aid in wholly understanding the effect of prolonged cortisol exposure on β-cell function in pregnancy.
We also noted that HCC was higher in overweight and obese women compared to those with normal weight during the prenatal period. Obesity is associated with insulin resistance, and changes in insulin sensitivity across pregnancy are partly related to changes in fat mass (42). Furthermore, obesity promotes the abnormal storage of lipids, inflammation, endothelial dysfunction, and consequently reduces placental metabolism and function (42). In late pregnancy, when cortisol reaches its maximum levels, associations were found with obesity among 768 pregnant women from the Ulm SPATZ Health Study in Germany, which is consistent with our findings using cortisol in mid-pregnancy (43). However, our BMI results differed from those of Bleker et al. (5). Some inconsistencies with the study by Bleker et al. (5) may arise from errors in the measurement of pre-pregnancy BMI, which was self-reported by them but directly measured in our study, and from differences in the sample used to assess cortisol (serum vs hair for us). Outside of pregnancy, hypercortisolism has been demonstrated to contribute to the association between impaired glucose metabolism, adiposity, obesity, and insulin resistance (7, 40).
Our study has several strengths. First, this is the first study looking at the influence of chronic stress, measured via HCC, on glucose tolerance in mid-pregnancy. Second, we used a relatively large sample size compared to prior studies and a good proportion of GDM cases (23%). This sample included deep phenotyping of pregnancy-related variables, allowing for multivariate adjustment for confounders. Furthermore, using HCC versus cortisol from blood, saliva, or urine, allowed us to assess maternal HPA activity and exposure to chronic stress for a prolonged period in pregnancy (3 months vs. 12-24 h). This study is not without some limitations. First, the cross-sectional analysis of HCC and maternal glycemia prevented us from investigating any longitudinal effect of HCC on glucose tolerance. Second, our study only included glycemic traits assessed in mid-pregnancy, and HCC was not available prenatally or in late pregnancy (third trimester). Third, despite the widely use of a 3-cm hair segment to represent retrospective cortisol levels over the past three months (26), more dynamic measures of cortisol using shorter hair segments might be important to examine (44). However, smaller segments (i.e., 1 cm) may impose a limitation as the threshold of minimum cortisol detection commonly falls below this length (44). Even though having samples from the same ethnic background helped us control for potential bias due to population stratification, validation of associations in a different population in pregnancy is required to support the generalizability of our evidence. Finally, although we have adjusted for covariates, the possibility of residual confounding due to unmeasured factors remains.
5. Conclusions
Using hair cortisol at two times in pregnancy and glycemic traits measured in mid-pregnancy, we demonstrated the hair cortisol's prospective and cross-sectional association with higher insulin resistance and increased β-cell function. The unexpected association of HCC with β-cell function may reflect an underlying compensatory mechanism prompted by lactogenic hormones to counteract the action of cortisol and maintain maternal glucose tolerance in pregnancy. Our study also demonstrated that cortisol did not directly influence the odds of GDM and was not associated with glucose levels in the OGTT. Future studies in larger samples and incorporating a longitudinal approach, will warrant a better understanding of the role of cortisol in glucose intolerance in pregnancy.
Supplementary Material
Highlights.
Hair cortisol is a well-known biomarker of chronic stress.
High maternal cortisol and glucose are associated with adverse pregnancy outcomes.
We investigated if cortisol influences glucose and insulin homeostasis in pregnancy.
Elevated hair cortisol is associated with insulin resistance in pregnancy.
Acknowledgments
The authors are grateful to the STEM-GDM study participants. The authors wish to thank the dedicated staff members of Asociacion Civil Proyectos en Salud (PROESA), Perú and Instituto Materno Perinatal, Perú, for their expert technical and administrative assistance with this research.
Role of Funding Source
This study was supported by Roche Diagnostic Operations Inc. (project number 208617-5074547) and the National Institute of Health (R21 HD102822). Sponsors had no role in study design; in the collection, analysis, and interpretation of data; in the writing of the manuscript.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Competing interests
The authors declare that they have no competing interests.
Declaration of interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Data sharing statement
Due to privacy/ethical restrictions, the data used in the development of this study is not publicly available, but access to it can be granted upon reasonable request to the corresponding author.
11. Reference list
- 1.Horan H, Cheyney M, Torres EG, Eick G, Bovbjerg M, Snodgrass JJ. Maternal hair cortisol concentrations across pregnancy and the early postpartum period in a Puerto Rican sample. Am J Hum Biol. 2022:e23718. [DOI] [PubMed] [Google Scholar]
- 2.Barker DJ. The origins of the developmental origins theory. J Intern Med. 2007;261(5):412–7. [DOI] [PubMed] [Google Scholar]
- 3.Traylor CS, Johnson JD, Kimmel MC, Manuck TA. Effects of psychological stress on adverse pregnancy outcomes and nonpharmacologic approaches for reduction: an expert review. Am J Obstet Gynecol MFM. 2020;2(4):100229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Barrett ES, Vitek W, Mbowe O, Thurston SW, Legro RS, Alvero R, et al. Allostatic load, a measure of chronic physiological stress, is associated with pregnancy outcomes, but not fertility, among women with unexplained infertility. Human Reproduction. 2018;33(9):1757–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bleker LS, Roseboom TJ, Vrijkotte TG, Reynolds RM, de Rooij SR. Determinants of cortisol during pregnancy - The ABCD cohort. Psychoneuroendocrinology. 2017;83:172–81. [DOI] [PubMed] [Google Scholar]
- 6.Duthie L, Reynolds RM. Changes in the maternal hypothalamic-pituitary-adrenal axis in pregnancy and postpartum: influences on maternal and fetal outcomes. Neuroendocrinology. 2013;98(2):106–15. [DOI] [PubMed] [Google Scholar]
- 7.Stalder T, Steudte-Schmiedgen S, Alexander N, Klucken T, Vater A, Wichmann S, et al. Stress-related and basic determinants of hair cortisol in humans: A meta-analysis. Psychoneuroendocrinology. 2017;77:261–74. [DOI] [PubMed] [Google Scholar]
- 8.Iob E, Steptoe A. Cardiovascular Disease and Hair Cortisol: a Novel Biomarker of Chronic Stress. Curr Cardiol Rep. 2019;21(10):116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.MENDELSON JH, STEIN S. Serum Cortisol Levels in Alcoholic and Nonalcoholic Subjects During Experimentally Induced Ethanol Intoxication. Psychosomatic Medicine. 1966;28(4):616–26. [DOI] [PubMed] [Google Scholar]
- 10.Steptoe A, Ussher M. Smoking, cortisol and nicotine. International Journal of Psychophysiology. 2006;59(3):228–35. [DOI] [PubMed] [Google Scholar]
- 11.Gerber M, Jonsdottir IH, Kalak N, Elliot C, Pühse U, Holsboer-Trachsler E, et al. Objectively assessed physical activity is associated with increased hair cortisol content in young adults. Stress. 2013;16(6):593–9. [DOI] [PubMed] [Google Scholar]
- 12.Gibson EL, Checkley S, Papadopoulos A, Poon L, Daley S, Wardle J. Increased Salivary Cortisol Reliably Induced by a Protein-Rich Midday Meal. Psychosomatic Medicine. 1999;61(2):214–24. [DOI] [PubMed] [Google Scholar]
- 13.Remer T, Maser-Gluth C, Wudy SA. Glucocorticoid measurements in health and disease--metabolic implications and the potential of 24-h urine analyses. Mini Rev Med Chem. 2008;8(2):153–70. [DOI] [PubMed] [Google Scholar]
- 14.Spiga F, Walker JJ, Terry JR, Lightman SL. HPA axis-rhythms. Compr Physiol. 2014;4(3):1273–98. [DOI] [PubMed] [Google Scholar]
- 15.Wright KD, Hickman R, Laudenslager ML. Hair Cortisol Analysis: A Promising Biomarker of HPA Activation in Older Adults. Gerontologist. 2015;55 Suppl 1(Suppl 1):S140–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hackett RA, Kivimäki M, Kumari M, Steptoe A. Diurnal Cortisol Patterns, Future Diabetes, and Impaired Glucose Metabolism in the Whitehall II Cohort Study. J Clin Endocrinol Metab. 2016;101(2):619–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Chiodini I, Adda G, Scillitani A, Coletti F, Morelli V, Di Lembo S, et al. Cortisol secretion in patients with type 2 diabetes: relationship with chronic complications. Diabetes Care. 2007;30(1):83–8. [DOI] [PubMed] [Google Scholar]
- 18.Joseph JJ, Wang X, Spanakis E, Seeman T, Wand G, Needham B, et al. Diurnal salivary cortisol, glycemia and insulin resistance: The multi-ethnic study of atherosclerosis. Psychoneuroendocrinology. 2015;62:327–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Zhuang W, Lv J, Liang Q, Chen W, Zhang S, Sun X. Adverse effects of gestational diabetes-related risk factors on pregnancy outcomes and intervention measures. Exp Ther Med. 2020;20(4):3361–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Diabetes IAo, Panel PSGC. International Association of Diabetes and Pregnancy Study Groups Recommendations on the Diagnosis and Classification of Hyperglycemia in Pregnancy. Diabetes Care. 2010;33(3):676–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Larrabure-Torrealva GT, Martinez S, Luque-Fernandez MA, Sanchez SE, Mascara PA, Ingar H, et al. Prevalence and risk factors of gestational diabetes mellitus: findings from a universal screening feasibility program in Lima, Peru. BMC Pregnancy and Childbirth. 2018;18(1):303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Metzger BE, Gabbe SG, Persson B, Buchanan TA, Catalano PA, Damm P, et al. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care. 2010;33(3):676–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412–9. [DOI] [PubMed] [Google Scholar]
- 24.Bonora E, Formentini G, Calcaterra F, Lombardi S, Marini F, Zenari L, et al. HOMA-Estimated Insulin Resistance Is an Independent Predictor of Cardiovascular Disease in Type 2 Diabetic Subjects: Prospective data from the Verona Diabetes Complications Study. Diabetes Care. 2002;25(7):1135–41. [DOI] [PubMed] [Google Scholar]
- 25.Orta OR, Tworoger SS, Terry KL, Coull BA, Gelaye B, Kirschbaum C, et al. Stress and hair cortisol concentrations from preconception to the third trimester. Stress. 2019;22(l):60–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Stalder T, Kirschbaum C. Analysis of cortisol in hair--state of the art and future directions. Brain Behav Immun. 2012;26(7): 1019–29. [DOI] [PubMed] [Google Scholar]
- 27.D'Anna-Hernandez KL, Ross RG, Natvig CL, Laudenslager ML. Hair cortisol levels as a retrospective marker of hypothalamic-pituitary axis activity throughout pregnancy: comparison to salivary cortisol. Physiol Behav. 2011;104(2):348–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Wennig R. Potential problems with the interpretation of hair analysis results. Forensic Science International. 2000;107(1):5–12. [DOI] [PubMed] [Google Scholar]
- 29.Gao W, Stalder T, Foley P, Rauh M, Deng H, Kirschbaum C. Quantitative analysis of steroid hormones in human hair using a column-switching LC–APCI–MS/MS assay. Journal of Chromatography B. 2013;928:1–8. [DOI] [PubMed] [Google Scholar]
- 30.van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate Imputation by Chained Equations in R. Journal of Statistical Software. 2011;45(3):1–67. [Google Scholar]
- 31.Kalra S, Einarson A, Karaskov T, Van Uum S, Koren G. The relationship between stress and hair cortisol in healthy pregnant women. Clin Invest Med. 2007;30(2):E103–7. [DOI] [PubMed] [Google Scholar]
- 32.Caparros-Gonzalez RA, Romero-Gonzalez B, Gonzalez-Perez R, Lucena-Prieto L, Perez-Garcia M, Cruz-Quintana F, et al. Maternal and Neonatal Hair Cortisol Levels Are Associated with Infant Neurodevelopment at Six Months of Age. J Clin Med. 2019;8(11):2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Duffy AR, Schminkey DL, Groer MW, Shelton M, Dutra S. Comparison of Hair Cortisol Levels and Perceived Stress in Mothers Who Deliver at Preterm and Term. Biol Res Nurs. 2018;20(3):292–9. [DOI] [PubMed] [Google Scholar]
- 34.Vejrazkova D, Vcelak J, Vankova M, Lukasova P, Bradnova O, Halkova T, et al. Steroids and insulin resistance in pregnancy. The Journal of Steroid Biochemistry and Molecular Biology. 2014;139:122–9. [DOI] [PubMed] [Google Scholar]
- 35.Arumugam R, Horowitz E, Lu D, Collier JJ, Ronnebaum S, Fleenor D, et al. The Interplay of Prolactin and the Glucocorticoids in the Regulation of β-Cell Gene Expression, Fatty Acid Oxidation, and Glucose-Stimulated Insulin Secretion: Implications for Carbohydrate Metabolism in Pregnancy. Endocrinology. 2008;149(11):5401–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Kirwan JP, Hauguel-De Mouzon S, Lepercq J, Challier J-C, Huston-Presley L, Friedman JE, et al. TNF-α Is a Predictor of Insulin Resistance in Human Pregnancy. Diabetes. 2002;51(7):2207–13. [DOI] [PubMed] [Google Scholar]
- 37.Feng Y, Feng Q, Qu H, Song X, Hu J, Xu X, et al. Stress adaptation is associated with insulin resistance in women with gestational diabetes mellitus. Nutrition & Diabetes. 2020;10(1):4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Keller-Wood M, Feng X, Wood CE, Richards E, Anthony RV, Dahl GE, et al. Elevated maternal cortisol leads to relative maternal hyperglycemia and increased stillbirth in ovine pregnancy. Am J Physiol Regul Integr Comp Physiol. 2014;307(4):R405–R13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Ullrich S, Berchtold S, Ranta F, Seebohm G, Henke G, Lupescu A, et al. Serum- and Glucocorticoid-Inducible Kinase 1 (SGK1) Mediates Glucocorticoid-Induced Inhibition of Insulin Secretion. Diabetes. 2005;54(4):1090–9. [DOI] [PubMed] [Google Scholar]
- 40.Ahmed SA, Shalayel MH. Role of cortisol in the deterioration of glucose tolerance in Sudanese pregnant women. East Afr Med J. 1999;76(8):465–7. [PubMed] [Google Scholar]
- 41.Kuo T, McQueen A, Chen TC, Wang JC. Regulation of Glucose Homeostasis by Glucocorticoids. Adv Exp Med Biol. 2015;872:99–126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Kampmann U, Knorr S, Fuglsang J, Ovesen P. Determinants of Maternal Insulin Resistance during Pregnancy: An Updated Overview. Journal of Diabetes Research. 2019;2019:5320156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Braig S, Grabher F, Ntomchukwu C, Reister F, Stalder T, Kirschbaum C, et al. Determinants of maternal hair cortisol concentrations at delivery reflecting the last trimester of pregnancy. Psychoneuroendocrinology. 2015;52:289–96. [DOI] [PubMed] [Google Scholar]
- 44.Maimon L, Milo T, Moyal RS, Mayo A, Danon T, Bren A, et al. Timescales of Human Hair Cortisol Dynamics. iScience. 2020;23(9):101501. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Due to privacy/ethical restrictions, the data used in the development of this study is not publicly available, but access to it can be granted upon reasonable request to the corresponding author.