Volume 11

Abstract

Background: Gestational diabetes mellitus (GDM) is prevalent in expecting mothers and is associated with various adverse maternal and fetal outcomes. Previous studies suggest a relationship between GDM and cesarean section (CS). However, the results of these studies in the context of primary CS and mothers’ age are limited and inconsistent, when other factors are accounted for.

Objective: This study compared primary CS risk between younger (15-34 years) and older (35+) mothers with GDM in Utah.

Methods: We conducted a cross-sectional analysis using the Utah Pregnancy Risk Assessment Monitoring System (PRAMS) phase 8 (2016-2021) survey data. A total of 8,491 responses were available for analysis. We estimated the risk/prevalence ratio (PR) of primary CS in women with GDM using Poisson regression models.

Results: 18.44% of women with GDM also had a primary CS, compared to 10.95% with no GDM. Overall, GDM significantly increased primary CS risk (PR 1.39: 1.13, 1.72). Younger mothers (15-34) with GDM also had a stronger adjusted risk (PR 1.48: 1.15, 1.91), but not older mothers (PR 1.40: 0.95, 2.07).

Conclusion: This study found greater primary CS risk in younger mothers with GDM but not older mothers. Longitudinal studies and interventions focusing on modifiable risk factors in younger Utah mothers are warranted to enhance their health during pregnancy and childbirth. More research is also needed to better understand CS decisions among younger Utah mothers, in particular, those who develop comorbidities such as GDM.

Implications: Culturally and demographically tailored interventions are needed to reduce GDM and associated primary CS risk in this population.

Introduction

Gestational Diabetes mellitus (GDM) is a common complication during pregnancy, affecting more than 14% of pregnancies globally and 9% of pregnancies in the United States (US).^1,2 In Utah, around 7.5% of pregnancies are affected by GDM.³ Some identified risk factors for GDM include advanced maternal age, obesity, and type 2 diabetes mellitus.⁴ Also, expecting mothers with GDM face an increased risk of preeclampsia, premature birth, cesarean section delivery (CS), and cardiovascular disease mortality.^5,6

Among the various complications linked to GDM, a major concern is its strong association with CS. In the US, CS is among the most common major surgical procedures performed, accounting for 32.1% of births in 2022.⁷ In Utah, CS prevalence was 26.3% among low risk mothers with no prior births, in the same year.⁸ This is noteworthy as CS is also associated with adverse maternal outcomes such as endomyometritis, uterine rupture, and death.^9,10

GDM is an independent predictor of CS in multiple previous studies.^11–14 However, there are direct contradictions surrounding the context of maternal age in the association between GDM and CS, when other factors are accounted for.^13,14 For example, one group found that the risk of CS was only increased in mothers aged 45 years or older.¹⁴ Other literature suggests a strong correlation of GDM and CS risk with increased age.¹³ Currently, there is no research investigating how maternal age may modify this association.

Generally, studies indicate age as a confounder, given that advanced maternal age has historically been linked to higher GDM and CS risks.¹³ However, recent reports suggest that younger mothers (under 35 years) may also face elevated risks of both conditions.^7,15 Also, there are no studies assessing differences between younger and older mothers with GDM in the risk of a first or primary CS; most existing research has focused on total CS, history of CS, or repeat CS,^11,13,16 requiring further investigation.

Moreover, to our knowledge, there has been no research examining age differences in primary CS risk within Utah mothers with GDM. To overcome these gaps, we utilized data from the Utah Pregnancy Risk Assessment Monitoring System (PRAMS) phase 8 survey data, which included 8,491 mothers. We analyzed the data by stratifying mothers with GDM into two groups: younger (15-34 years) and older (35+), to assess their risk of a primary CS.

Methods

Study Design and Population

This study utilized the Utah Pregnancy Risk Assessment Monitoring System (PRAMS) phase 8 (2016-2021) survey data for analysis. PRAMS is an ongoing population-based survey designed to collect information on maternal behaviors and experiences before, during, and immediately following pregnancy.¹⁷ Approximately 200 new mothers are randomly selected monthly for participation using Utah Birth Certificates. All responses are linked to the birth certificate record throughout the data collection process. Responses to the PRAMS survey are also weighted to accurately represent all women given birth in Utah.¹⁷ The dataset for the present study included 8,491 women, which reflected an extrapolated population size of 279,355 women in Utah (after weighing) using SAS analytical software package.

Exposure

The primary exposure was GDM, categorized as “Yes” or “No.” The information on GDM status was obtained from the PRAMS dataset which was linked to the infant’s birth certificate record.

Outcome

The outcome of interest was primary CS delivery categorized as “Yes” or “No.” The information on CS delivery was obtained from the PRAMS dataset which was linked to the infant’s birth certificate record.A “yes” for primary CS delivery indicates women with their first CS delivery reported in the birth certificate records, irrespective of the number of previous live births.

Covariates

Covariates used in the study included age (<20, 20-24, 25-34, and 35+ years of age), race (white or non-white), ethnicity (Hispanic or non-Hispanic), degree attained (associate’s degree or lower, bachelor’s degrees or higher, and unknown degree), region of residence (urban or rural), body mass index BMI (underweight: <18.5 kg/m²; healthy weight: 18.5-24.9 kg/m²; overweight: 25-29.9 kg/m²; and obese: >30 kg/m²),¹⁸ hypertension (yes or no), number of previous live births (0, 1-2, 3-4, and 5 or more), and infant birth weight (grams).

Statistical Analysis

All analyses were conducted using SAS Studio (version 9.4). Sampling weights were applied to the analysis to obtain representative Utah population estimates. The PRAMS weighing methodology was provided by the CDC and Utah Department of Health and Human Services (UDHHS).¹⁹ Group differences for sample characteristics were calculated using chi-squared tests for categorical variables and student t-tests for continuous variables. The frequency distributions by GDM and primary CS are presented in Table 1 and eTables 1 and 2 (see Supplementary Materials).

The modified Poisson regression was used to examine weighted crude and adjusted associations between GDM and CS.²⁰ Prevalence Ratio (PR) estimates of the regression analysis are presented in Table 2. The Poisson regression was also used for age-stratified analysis of primary CS risk. Individual estimates for younger (15-34 years) and older mothers (35+) are also in Table 2. Age 35 and older was used as the cutoff point for stratification due to research suggesting 35 years as advanced maternal age.²¹

Ethics Approval

The collection of the data used in this study was approved by Utah Department of Health and Human Services (UDHHS) institutional review board.

Results

Population Characteristics

The study consisted of 8,491 respondents, reflecting 279,355 women giving birth in Utah after weighting. Mean (standard error) age was 28.68 (0.07) years, 89.45% identified as White race, and 83.73% identified as non-Hispanic ethnicity (Tables 1 and 2). In addition, 76.48% of women in the study lived in an urban setting, and 59.78% held an associate degree or lower.

About 6.70% of mothers in the study were diagnosed with GDM, and 11.45% of mothers had a primary CS. Mothers with GDM were more often older, evidenced by a mean age of 31.16 (standard error; SE = 0.22) years compared to 28.51 (SE = 0.06) years among mothers without GDM (P<.0001) (Table 1). Differences were observed between mothers with GDM and those without GDM in terms of race (P<.0001), ethnicity (P=0.0026), degree attainment (P=0.0051), BMI (P<.0001), comorbidity of hypertension (P<.0001), and number of previous live births (P=0.0030). (Table 1). Mothers with primary CS were more often Hispanic ethnicity (P=0.0038), had a higher BMI (P<.0001), comorbidity of hypertension (P<.0001), no previous live births (P<.0001), and lower infant birth weight (P<.0001).

Regarding the distribution of primary CS by GDM, 18.44% of mothers with GDM had primary CS, while 10.95% of mothers without GDM had a primary CS (P<.0001) (Table 1). Further, among younger mothers (15-34 years) with GDM, 25.47% had a primary CS. In contrast, among older mothers (35+) with GDM, 18.32% had a primary CS (Table 2).

Association between GDM and Primary CS delivery

Overall, GDM was strongly associated with primary CS (Table 2). Mothers with GDM had a 1.39 times higher adjusted risk of primary CS compared to those without GDM (PR 1.39: 1.13, 1.72). In the age stratified analysis, the adjusted risk of primary CS was 1.48 times significantly higher for mothers aged 15-34  (PR 1.48: 1.15, 1.91). In contrast, the adjusted risk for mothers aged 35+ years was (PR 1.40: 0.95, 2.07).

Discussion

Principal Findings

The present study provides a first look at primary CS risk in Utah mothers with GDM. It is unique in that it investigated differences between younger (15-34 years) and older (35+) mothers. We found that Utah mothers with GDM had a significantly higher risk of primary CS, and observed a stronger association in younger mothers (15-34 years).

Interpretation

Our first major finding regarding a higher risk of primary CS in mothers with GDM is consistent with other research.²³ For example, the 2024 study by Fresch et al.²³ found a 1.34 times higher adjusted risk of primary CS in mothers with GDM (PR 1.34: 1.31, 1.36). Medical conditions like suspected macrosomia, labor arrest, and indeterminate fetal heart rate tracing may make vaginal birth difficult and unsafe, requiring a CS.^16,24 Importantly, GDM is a known independent risk factor which can exacerbate the occurrence of these conditions.^24–26

In our second major finding, the risk of primary CS in mothers with GDM was higher across age groups but only significant for younger mothers ages 15-34. The estimates for this age group (PR 1.48: 1.15, 1.91) were higher than those of the overall population (PR 1.39: 1.13, 1.72). This result is surprising given that research generally indicates greater GDM and primary CS risk in advanced or older maternal age only.²⁷ It is important to note that existing research has not specifically examined the associations in younger mothers. Instead, these studies have only used younger age as a comparison or reference group in their analyses.^23,27 A recent US-based study, however, comparing primary CS prevalence between 2021 and 2022 showed increases for mothers ages 25-29 (21.0% vs 21.1%), 30-34 (33.8% vs 33.9%), 35-39 (26.1% vs 26.4%), and 40+ (33.7% vs 34.0%).^7,28 The upward trend in younger mothers is noteworthy and highlights two key considerations. First, the risk of pregnancy complications in this age group may be higher than current estimates indicate. Second, younger expecting mothers may be underrepresented in maternity health research.

Several factors may be contributing to the rising CS rates among younger mothers. Common reasons include fear of labor pain (as a first time mother or following a previous birth), concerns about bodily changes and damage, concerns about their baby’s health, and convenience of scheduling a CS birth.^29,30 In Utah, a 2013 PRAMS data report revealed labor challenges indicated by fetal monitor, prolonged labor, and failed labor induction as reasons for CS in mothers reporting a primary CS.³¹ Given that GDM is associated with many, if not all, of these complications, it is possible that a diagnosis of the condition may heighten fear of a vaginal delivery, particularly among younger expecting mothers, which may result in preference for a CS. We recommend further research be performed to confirm this relationship.

Health Implications

CS intersects with various domains of health. In terms of physical health, CS has associated complications for both the mother and child. CS is a risk factor for common gynecological conditions, including the development of urinary tract infections, gastrointestinal problems, adhesions, pain, infertility, painful menses, and endometriosis in the mother.^9,32 Additionally, children born via CS are at increased risk for respiratory tract infections, asthma, and obesity.³³

The economic health domain is also impacted. On average, CS is more costly for the patient and for the hospital than vaginal delivery. Across Utah, total CS costs are reported as $8,952.52 for those with insurance and $14,252.80 for those without insurance while vaginal births are reported as $5,951.76 and $10,199.52 with and without insurance, respectively.³⁴ Much of this cost is taken care of by insurance companies. However, these costs do not consider any complications stemming from delivery or care for mother or child. For the social health domain, CS can impact family planning and relationships due to potential complications such as painful intercourse and infertility.^9,32,35 In mental and behavioral health, the increased risk of complications and the recovery process resulting from CS can lead to stress, anxiety, and even delayed initiation of breastfeeding.^36,37

In 2023, Utah’s CS rate (19.4%) was lower than the average rate across the United States (US, 26.3%) for low-risk women without prior births.⁸ This trend has remained consistent over time (2013-2023).⁸ Additionally, Utah has a lower rate of GDM (7.5%) compared to the US (9%), although this difference is small.³ One explanation for this difference is the younger age of mothers at their first birth in Utah compared to the US (25.9 in Utah versus 27.1 years overall in 2020).³⁸

It is critical that the aforementioned issues are addressed to improve maternal and child health outcomes. We recommend the following actions for health systems and population health researchers in Utah. First, the necessity of CS should be carefully considered in the context of each expecting mother. It is equally important that cases of CS in young expecting mothers with GDM are closely reviewed to ensure that decisions for CS follow appropriate clinical guidelines. Second, health systems should provide interdisciplinary team-based care to support young mothers who develop comorbidities (such as GDM) to mitigate fear and worry about their pregnancy. Third, significant outreach is needed to increase awareness about comorbidities during pregnancy and how expecting mothers can protect their health and the health of their baby. Lastly, more work is needed in Utah to ensure that new and expecting younger mothers are represented in maternity health research.

Conclusions

In summary, this study contributes valuable insights into the association between GDM and primary CS in Utah, highlighting the importance of early detection and management of GDM to optimize maternal and infant health outcomes. Longitudinal studies and interventions focusing on modifiable risk factors in younger Utah mothers are warranted to enhance their health during pregnancy and childbirth. More research is also needed to better understand CS decisions among younger Utah mothers, in particular, those who develop comorbidities such as GDM.

Strengths & Limitations

This study had several strengths and limitations. First, our weighted sample was representative of the Utah population (279,355). This enabled more precise estimates and comprehensive evaluation of overall findings. Second, this study was unique as it is the first to examine GDM and the risk of CS within the Utah population. Concerning limitations, since the study utilized Utah population data, the findings may not be generalizable to other areas. Further, the study design was cross-sectional, which introduces temporal ambiguity. As this analysis was done at one point, it cannot establish a cause-and-effect relationship.

Acknowledgements

Data was provided by the Utah Pregnancy Risk Assessment and Monitoring System (PRAMS), a project of the UDHHS, Office of Maternal and Child Health (MCH). No funding was provided for the project. This information or content and conclusions are those of the author and should not be construed as the official position or policy of, nor should any endorsements be inferred by UDHHS MCH.

References

1.   Wang H, Li N, Chivese T, et al. IDF Diabetes Atlas: Estimation of Global and Regional Gestational Diabetes Mellitus Prevalence for 2021 by International Association of Diabetes in Pregnancy Study Group’s Criteria. Diabetes Res Clin Pract. 2022;183:109050. doi:10.1016/j.diabres.2021.109050

2.   Center for Disease Control and Prevention. About Gestational Diabetes. Diabetes. Published May 31, 2024. Accessed July 25, 2024. https://www.cdc.gov/diabetes/about/gestational-diabetes.html

3.   Utah Department of Health and Human Services. IBIS-PH – Health Indicator Report – Diabetes: gestational diabetes. Accessed July 24, 2024. https://ibis.utah.gov/ibisph-view/indicator/view/DiabGestDiab.html

4.   Casagrande SS, Linder B, Cowie CC. Prevalence of gestational diabetes and subsequent Type 2 diabetes among U.S. women. Diabetes Res Clin Pract. 2018;141:200-208. doi:10.1016/j.diabres.2018.05.010

5.   Preda A, Pădureanu V, Moța M, et al. Analysis of Maternal and Neonatal Complications in a Group of Patients with Gestational Diabetes Mellitus. Medicina (Kaunas). 2021;57(11):1170. doi:10.3390/medicina57111170

6.   Wang YX, Mitsunami M, Manson JE, et al. Association of Gestational Diabetes With Subsequent Long-Term Risk of Mortality. JAMA Internal Medicine. 2023;183(11):1204-1213. doi:10.1001/jamainternmed.2023.4401

7.   Osterman M, Hamilton B, Martin J, Driscoll A. National Vital Statistics Reports Volume 73, Number 2, April 4, 2024. Published online April 4, 2024.

8.   Utah Department of Health and Human Services. IBIS-PH – Health Indicator Report – Cesarean delivery among low risk women with no prior births. Accessed July 24, 2024. https://ibis.utah.gov/ibisph-view/indicator/view/CesDel.UT_US.html

9.   Quinlan JD, Murphy NJ. Cesarean Delivery: Counseling Issues and Complication Management. afp. 2015;91(3):178-184.

10. Boyle A, Reddy UM, Landy HJ, Huang CC, Driggers RW, Laughon SK. Primary Cesarean Delivery in the United States. Obstet Gynecol. 2013;122(1):33-40. doi:10.1097/AOG.0b013e3182952242

11. Gorgal R, Gonçalves E, Barros M, et al. Gestational diabetes mellitus: A risk factor for non-elective cesarean section. Journal of Obstetrics and Gynaecology Research. 2012;38(1):154-159. doi:10.1111/j.1447-0756.2011.01659.x

12. Song J, Cai R. Interaction between smoking during pregnancy and gestational diabetes mellitus and the risk of cesarean delivery: evidence from the National Vital Statistics System 2019. The Journal of Maternal-Fetal & Neonatal Medicine. 2023;36(2):2259048. doi:10.1080/14767058.2023.2259048

13. Akinyemi OA, Weldeslase TA, Odusanya E, et al. Profiles and Outcomes of Women with Gestational Diabetes Mellitus in the United States. Cureus. 2023;15(7):e41360. doi:10.7759/cureus.41360

14. Claramonte Nieto M, Meler Barrabes E, Garcia Martínez S, Gutiérrez Prat M, Serra Zantop B. Impact of aging on obstetric outcomes: defining advanced maternal age in Barcelona. BMC Pregnancy and Childbirth. 2019;19(1):342. doi:10.1186/s12884-019-2415-3

15. Shah NS, Wang MC, Freaney PM, et al. Trends in Gestational Diabetes at First Live Birth by Race and Ethnicity in the US, 2011-2019. JAMA. 2021;326(7):660-669. doi:10.1001/jama.2021.7217

16. Caughey AB, Cahill AG, Guise JM, Rouse DJ. Safe prevention of the primary cesarean delivery. American Journal of Obstetrics & Gynecology. 2014;210(3):179-193. doi:10.1016/j.ajog.2014.01.026

17. Utah Department of Health and Human Services. Utah PRAMS | Maternal and Infant Health Program. Accessed July 25, 2024. https://mihp.utah.gov/pregnancy-and-risk-assessment

18. CDC. Defining Adult Overweight and Obesity. Centers for Disease Control and Prevention. Published June 3, 2022. Accessed September 28, 2022. https://www.cdc.gov/obesity/basics/adult-defining.html

19. Center for Disease Control and Prevention. Data Methodology. Pregnancy Risk Assessment Monitoring System (PRAMS). Published May 20, 2024. Accessed July 25, 2024. https://www.cdc.gov/prams/php/methodology/index.html

20. Lindquist K. How can I estimate relative risk in SAS using proc genmod for common outcomes in cohort studies? | SAS FAQ. Accessed January 17, 2024. https://stats.oarc.ucla.edu/sas/faq/how-can-i-estimate-relative-risk-in-sas-using-proc-genmod-for-common-outcomes-in-cohort-studies/

21. American College of Obstetricians and Gynecologists (ACOG). Pregnancy at Age 35 Years or Older: ACOG Obstetric Care Consensus No. 11. Obstetrics & Gynecology. 2022;140(2):348. doi:10.1097/AOG.0000000000004873

22. Marshall SW. Power for tests of interaction: effect of raising the Type I error rate. Epidemiol Perspect Innov. 2007;4:4. doi:10.1186/1742-5573-4-4

23. Fresch R, Stephens K, DeFranco E. The Combined Influence of Maternal Medical Conditions on the Risk of Primary Cesarean Delivery. AJP Rep. 2024;14(1):e51-e56. doi:10.1055/s-0043-1777996

24. Kc K, Shakya S, Zhang H. Gestational diabetes mellitus and macrosomia: a literature review. Ann Nutr Metab. 2015;66 Suppl 2:14-20. doi:10.1159/000371628

25. Gill P, Henning JM, Carlson K, Van Hook JW. Abnormal Labor. In: StatPearls. StatPearls Publishing; 2024. Accessed July 26, 2024. http://www.ncbi.nlm.nih.gov/books/NBK459260/

26. Depla AL, De Wit L, Steenhuis TJ, et al. Effect of maternal diabetes on fetal heart function on echocardiography: systematic review and meta‐analysis. Ultrasound Obstet Gynecol. 2021;57(4):539-550. doi:10.1002/uog.22163

27. Richards MK, Flanagan MR, Littman AJ, Burke AK, Callegari LS. Primary cesarean section and adverse delivery outcomes among women of very advanced maternal age. J Perinatol. 2016;36(4):272-277. doi:10.1038/jp.2015.204

28. Osterman MJK, Hamilton BE, Martin JA, Driscoll AK, Valenzuela CP. Births: Final Data for 2021. Natl Vital Stat Rep. 2023;72(1):1-53.

29. Stoll KH, Hauck YL, Downe S, Payne D, Hall WA. Preference for cesarean section in young nulligravid women in eight OECD countries and implications for reproductive health education. Reprod Health. 2017;14:116. doi:10.1186/s12978-017-0354-x

30. Stoll K, Edmonds JK, Hall WA. Fear of Childbirth and Preference for Cesarean Delivery Among Young American Women Before Childbirth: A Survey Study. Birth. 2015;42(3):270-276. doi:10.1111/birt.12178

31. Utah Department of Health and Human Services. Utah Health Status Update: Primary Cesarean Delivery. Published June 2013. Accessed July 26, 2024. https://ibis.utah.gov/ibisph-view/pdf/opha/publication/hsu/2013/1306_CSection.pdf

32. Antoine C, Young BK. Cesarean section one hundred years 1920-2020: the Good, the Bad and the Ugly. J Perinat Med. 2020;49(1):5-16. doi:10.1515/jpm-2020-0305

33. Słabuszewska-Jóźwiak A, Szymański JK, Ciebiera M, Sarecka-Hujar B, Jakiel G. Pediatrics Consequences of Caesarean Section-A Systematic Review and Meta-Analysis. Int J Environ Res Public Health. 2020;17(21):8031. doi:10.3390/ijerph17218031

34. Peter K. Costs of Childbirth by State. PolicyScout. August 17, 2022. Accessed August 15, 2024. https://policyscout.com/health-insurance/learn/costs-childbirth-by-state#

35. Kainu JP, Sarvela J, Tiippana E, Halmesmäki E, Korttila KT. Persistent pain after caesarean section and vaginal birth: a cohort study. Int J Obstet Anesth. 2010;19(1):4-9. doi:10.1016/j.ijoa.2009.03.013

36. Skov SK, Hjorth S, Kirkegaard H, Olsen J, Nohr EA. Mode of delivery and short-term maternal mental health: A follow-up study in the Danish National Birth Cohort. Int J Gynaecol Obstet. 2022;159(2):457-465. doi:10.1002/ijgo.14155

37. Hosaini S, Yazdkhasti M, Moafi Ghafari F, Mohamadi F, Kamran Rad SHR, Mahmoodi Z. The relationships of spiritual health, pregnancy worries and stress and perceived social support with childbirth fear and experience: A path analysis. PLoS One. 2023;18(12):e0294910. doi:10.1371/journal.pone.0294910

38. Pieper K, Seager M, Blackburn R. Utah Women and Fertility: Trends and Changes from 1970–2021. Published online April 4, 2023.

Supplementary Materials

Citation

Adediran E, Duffy HR, & Asay KM. (2024). Does Maternal Age Modify the Association Between Gestational Diabetes and Primary C-section Delivery? A Cross-Sectional Study of Utah PRAMS Data. Utah Women’s Health Review. doi: 10.26054/d-yx18-mp3d

PDF

View/download

Abstract

Objectives: The objective of this study is to test the association between preconception polycystic ovary syndrome (PCOS) and gestational diabetes mellitus (GDM) using Utah’s Pregnancy Risk Assessment Monitoring System (2016-2021). In addition, pre-pregnancy hypertension will be tested as a potential effect moderator.

Methods: This cross-sectional study utilizes data from Phase 8 of the Utah Pregnancy Risk Assessment Monitoring System (PRAMS) survey (2016-2021). The association between PCOS and GDM was tested using Poisson regression to generate adjusted prevalence ratios and 95% confidence intervals.

Results: PCOS was associated with higher prevalence of GDM in all models, regardless of whether the outcome data (GDM) came from the infant’s birth certificate, the PRAMS survey, or the combined measure. When adjusting for sociodemographic characteristics, lifestyle factors, reproductive history, and comorbidities, women with PCOS were 1.50 (1.16-1.95) times as likely to have GDM (reported on birth certificate and/or survey) compared to women without PCOS. Pre-pregnancy hypertension was not found to be a statistically significant effect moderator.

Conclusion: The findings from this study were consistent with the majority of research indicating that women with PCOS have increased risk for GDM. This is also the first known study to test pre-pregnancy hypertension as an effect moderator between PCOS and GDM. More research is needed on the role of comorbidities such as chronic hypertension as effect modifiers between PCOS and GDM.

Implications: These findings show that women with PCOS are at high risk for GDM, among a population-based sample of mothers. Interventions to reduce the risk of GDM among women with PCOS need to be developed and evaluated.

Introduction

An estimated 5-15% of women have polycystic ovary syndrome (PCOS).¹ The three diagnostic features according to the Rotterdam Criteria are ovulatory dysfunction such as anovulation and oligo-ovulation, hyperandrogenism, and polycystic ovaries. Patients must display two of the three to be diagnosed.² In addition, clinicians must exclude other conditions with similar presentations, such as congenital adrenal hyperplasia, Cushing’s syndrome and androgen-secreting tumors, and drug-induced androgen excess.²

PCOS is a major contributor to infertility. Ovulatory disorders account for approximately 25% of infertility diagnoses; 70% of women with anovulation have polycystic ovary syndrome.³ In addition to being at increased risk for infertility, women with PCOS have elevated risk of obesity, type II diabetes, hyperlipidemia, depression, anxiety, obstructive sleep apnea, nonalcoholic fatty liver disease, endometrial cancer, and cardiovascular disease.^4,5 The cause of PCOS is unknown and there is no cure.

Gestational diabetes mellitus (GDM) is a glucose tolerance disorder that starts during pregnancy and affects approximately 6% of pregnancies. The prevalence of GDM has had a relative increase of approximately 78% in the last two decades. The rising prevalence of GDM over the last few decades is concerning because GDM is associated with pre-eclampsia, macrosomia, and future metabolic diseases in the mother and child.⁶

Prior studies have produced mixed findings regarding the presence and magnitude of the association between PCOS and GDM. The majority of the literature including the most recent meta-analysis found that women with PCOS have elevated risk of GDM compared to women without PCOS.⁶ The objective of this study is to test the association between preconception PCOS and GDM using Utah’s Pregnancy Risk Assessment Monitoring System (2016-2021). Utilizing these data allows the association between preconception PCOS and GDM to be tested among a representative population-based sample. This study includes women who are often underrepresented in hospital or clinic-based studies, such as Hispanics and women of low socioeconomic status.

Another unique feature of this study is that pre-pregnancy hypertension will be tested as a potential effect moderator between PCOS and GDM. PCOS is considered a heterogeneous disorder with multiple phenotypes which affect a person’s health risks at different stages of life.⁷ The purpose of testing pre-pregnancy hypertension as an effect moderator is to determine if the risk of GDM differs among women with and without pre-pregnancy hypertension. This information could contribute to our understanding of the underlying mechanism between PCOS and GDM as well as inform clinical decisions and treatment.⁷

Methods

Study Design

This cross-sectional study utilizes data from the 2016-2021 Utah Pregnancy Risk Assessment Monitoring System (PRAMS) survey which is a joint project between state, local, tribal, and territorial health departments and the Centers for Disease Control and Prevention (CDC).

Data Sources

PRAMS data collection began in 1987 to understand why some infants are born healthy and others are not. It was designed to identify groups of women and infants at high risk for health problems, monitor changes in health status, and measure progress toward goals in improving the health of mothers and infants.⁸ PRAMS surveillance began in Utah in 1999.⁹

Approximately 200 new mothers are randomly selected to participate each month from Utah birth certificates.⁹ Recruitment for PRAMS occurs according to the protocol developed by the CDC which combines statewide mailings of the surveys and telephone follow-up to women who do not complete the survey by mail. A key feature of PRAMS is the stratified systematic sampling, which over samples from populations of interest based on factors such as maternal age, race/ethnicity, geographic area of residence, and infant birth weight.¹⁰ The survey is available in both English and Spanish.

The data used in this study were from the Utah PRAMS Phase 8 (2016-2021) questionnaire (n=8,491) reflecting an estimated population of 279,355 women. The design and sampling frame assure the study sample is representative of Utah’s population. The unweighted response rates in Utah were 55.6% in 2016, 60.2% in 2017, 54% in 2018, 69.5% in 2019, 60.9% in 2020, and 52.6% in 2021.⁸

Outcome Measures

The outcome of interest for this study was whether or not a woman was diagnosed with gestational diabetes mellitus (GDM) during her most recent pregnancy. This outcome was measured in two ways. First, as a field on the birth certificate and second as a self-reported item on the PRAMS survey. On the survey women were asked, “during your most recent pregnancy, did you have any of the following health conditions?” One of the listed conditions was, “gestational diabetes (diabetes that started during this pregnancy)” and respondents were instructed to check yes or no. Estimates of the association between preconception polycystic ovary syndrome (PCOS) and gestational diabetes mellitus (GDM) will be provided using each outcome measurement method separately as well as a combined measure, which includes mothers who self-reported GDM for index pregnancy via PRAMS survey and those for whom GDM was indicated on the birth certificate. 8012 observations (96.4%) were categorized consistently for GDM between the survey and the birth certificate. However, 85 respondents did not self-report whether or not they had GDM on the survey and 59 respondents were categorized as not having GDM on the survey though it was indicated on the birth certificate. Conversely, 158 respondents self-reported having GDM on the survey though it was not indicated on the birth certificate. The combined GDM measure is the most inclusive outcome measurement and classifies observations as having had GDM if they self-reported GDM on the survey or if GDM was indicated on the birth certificate.

In prior studies, both birth certificate data and maternal recall have demonstrated excellent validity for variables related to labor and delivery when compared to medical records.¹¹ Furthermore, one study using New York state PRAMS data found good agreement between birth certificate and self-reporting on the PRAMS survey specifically for GDM.¹² Though both measurement methods are believed to have a high degree of validity, due to the 85 missing values for GDM on the survey, the birth certificate record may be the most valid measurement method for this study.

Covariates

Potential confounding factors believed to influence both PCOS and GDM were determined based on prior literature.^1,4,5,6 Sociodemographic covariates include maternal age in years, yearly total household income before taxes during the 12 months before the new baby was born (categorical), maternal education (categorical), race/ethnicity (white/non-white and Hispanic/non-Hispanic), marital status (married/not married), and urban/rural residence. Salt Lake, Davis, Utah, and Weber counties were categorized as urban, and all others were categorized as rural. Lifestyle factors included cigarette smoking within the last two years (yes/no), drinking within the last two years (yes/no), and pre-pregnancy BMI (categorical). Reproductive history and comorbidities including previous live birth (yes/no), infertility treatment (yes/no), depression during the three months before getting pregnant (yes/no), and anxiety during the three months before getting pregnant (yes/no) were also considered as potential confounding factors. The following covariates were collected on the survey: maternal age, family income, cigarette smoking, drinking, pre-pregnancy BMI, pre-pregnancy depression, and pre-pregnancy anxiety. The following covariates were collected on the birth certificate: maternal education, race/ethnicity, marital status, urban/rural residence, previous live birth, and infertility treatment.

Respondents self-reported hypertension during the three months before getting pregnant (yes/no) on the survey. Pre-pregnancy hypertension was tested as an effect moderator is to determine if the risk of GDM differs among women with PCOS with and without comorbid hypertension.

Statistical Analysis

After excluding observations with missing values for PCOS (n=156) and hypertension (n=21), 8314 women (97.9%) were included in the analyses, reflecting an estimated population size of 274,023 women (Figure 1). Participant characteristics were reported by PCOS status, the exposure of interest. All variables were categorical and are reported as weighted percentages. Descriptive characteristics of respondents with and without PCOS were calculated by chi-square tests, taking into account the stratified survey sampling.

The association between PCOS and gestational diabetes was tested using Poisson regression models with a robust error variance. The models accounted for PRAMS’ use of stratified sampling and generated adjusted prevalence ratios (PR) and 95% confidence intervals (CI). The reference group in all models was mothers without PCOS. Pre-pregnancy hypertension was tested as an effect modifier using a stratified analysis and the Wald test. SAS Studio 9.4 and Stata 16.1 were used for data analysis.

Results

Characteristics of the Mothers

The majority of the respondents were non-Hispanic white (77.7%), received education beyond high school (70.4%), were married (81.7%), and lived in an urban area (76.4%) (Table 1). The median maternal age was 28.5 among respondents with PCOS and 27.9 among respondents without PCOS (Table 1). Most of the women were parous (65.4%), had a normal pre-pregnancy BMI (47.9%), and had not smoked cigarettes (90.7%) or drank alcohol (65.6%) within the last two years before becoming pregnant (Table 1).

^{Table 1. Sociodemographic, lifestyle and clinical characteristics of women by polycystic ovary syndrome (PCOS) status, Utah PRAMS, 2016-2021, n=8,314, reflecting an estimated population size of 274,023 women}

Characteristic	Overall	No PCOS	PCOS	P-value (chi-square)
Maternal Age (%)
15-19	3.2	3.3	1.4
20-29	54.3	54.4	52.7
30-39	40.0	39.7	43.0
40+	2.6	2.6	3.0	0.08
Family Income (%)
$0-$20,000	14.2	14.5	11.3
$20,001-$32,000	13.5	13.7	10.7
$32,001-$57,000	22.1	22.1	22.3
$57,001-$85,000	20.9	20.8	22.3
$85,001+	23.6	23.1	29.4	<0.00
Maternal Education (%)
Less than HS	7.4	7.7	4.3
HS/GED	19.6	19.7	17.4
Some College	21.8	21.8	22.0
Associate	10.7	10.5	13.1
Bachelors	30.1	29.9	32.1
Masters, Doctorate/Professional	7.8	7.7	8.6	0.09
Race/ethnicity (%)
Hispanic, non-white	4.2	4.4	2.0
Hispanic, white	11.9	12.0	11.1
Non-Hispanic, white	77.7	77.3	82.5
Non-Hispanic, non-white	6.0	6.2	4.1	0.01
Married (%)	81.7	81.5	84.5	0.13
Rural (%)	23.6	23.7	21.4	0.30
Smoking (%)	9.3	9.4	9.3	0.07
Drinking (%)	34.4	34.2	36.4	0.52
Body Mass Index (BMI) (%)
Underweight (<18.5)	3.8	4.0	1.7
Normal (18.5-24.9)	47.9	49.2	31.3
Overweight (25.0-29.9)	24.2	24.3	22.7
Obese (30.0+)	22.9	21.2	42.8
Unknown	1.3	1.3	1.5	<0.00
Depression (%)	18.5	17.1	35.9	<0.00
Anxiety (%)	28.1	26.4	48.8	<0.00
Hypertension (%)	3.9	2.4	22.5	<0.00
Previous live birth (%)	65.4	65.7	62.6	0.21
Infertility treatment (%)	6.8	5.0	30.1	<0.00

Compared to women without PCOS, women with PCOS were more likely to be non-Hispanic white, have higher family income, be older, and have higher educational attainment (Table 1). Women with PCOS also had significantly higher rates of obesity (42.8% vs. 21.2%), depression (35.9% vs. 17.1%), anxiety (48.8% vs. 26.4%), hypertension (22.5% vs. 2.4%), and utilization of infertility treatment (30.1% vs. 5.0%) compared to women without PCOS (Table 1).

Association Between Polycystic Ovary Syndrome (PCOS) and Gestational Diabetes Mellitus (GDM)

PCOS was associated with higher prevalence of GDM in all models, regardless of whether the outcome data came from the infant’s birth certificate, the PRAMS survey, or the combined measure. Estimates did not meaningfully differ between the different GDM variables (Table 2). After adjusting for sociodemographic characteristics of age, educational attainment, family income, race/ethnicity, marital status, and urban/rural, women with PCOS were 2.14 times (95% CI: 1.63–2.81) as likely to have GDM using the birth certificate measure, 2.0 times (95% CI: 1.54–2.60) as likely to have GDM using the PRAMS survey measure, and 2.02 times (95% CI: 1.57–2.59) as likely to have GDM using the combined measure compared to women without PCOS (Table 2).

After further adjustment by lifestyle factors of cigarette smoking, alcohol drinking, and BMI, women with PCOS were 1.64 times (95% CI: 1.24–2.17) as likely to have GDM using the birth certificate measure, 1.56 times (95% CI: 1.20–2.04) as likely to have GDM using the PRAMS survey measure, and 1.58 times (95% CI: 1.23–2.03) as likely to have GDM using the combined measure compared to women without PCOS (Table 2).

The final model was further adjusted by reproductive history and comorbidities including previous live birth, infertility treatment, depression, anxiety, and hypertension. In the final model women with PCOS were 1.49 (95% CI: 1.11–2.01) times as likely to have GDM using the birth certificate measure, 1.53 times (95% CI: 1.16–2.01) times as likely to have GDM using the PRAMS survey measure, and 1.50 (95% CI 1.16–1.95) times as likely to have GDM using the combined measure compared to women without PCOS (Table 2).

^{Table 2. Association between polycystic ovary syndrome (PCOS) and gestational diabetes mellitus (GDM)}

Gestational Diabetes (GDM)
	Unadjusted PR (95% CI)	Model 1 PR (95% CI)	Model 2 PR (95% CI)	Model 3 PR (95% CI)
GDM – Birth Certificate
PCOS	2.04 (1.56-2.68)	2.14 (1.63-2.81)	1.64 (1.24-2.17)	1.49 (1.11-2.01)
No PCOS	1.0 (Reference)	1.0 (Reference)	1.0 (Reference)	1.0 (Reference)
GDM – Survey
PCOS	1.94 (1.51-2.51)	2.00 (1.54-2.60)	1.56 (1.20-2.04)	1.53 (1.16-2.01)
No PCOS	1.0 (Reference)	1.0 (Reference)	1.0 (Reference)	1.0 (Reference)
GDM – Birth Certificate or Survey
PCOS	1.92 (1.50-2.45)	2.02 (1.57-2.59)	1.58 (1.23-2.03)	1.50 (1.16-1.95)
No PCOS	1.0 (Reference)	1.0 (Reference)	1.0 (Reference)	1.0 (Reference)

Association Between Polycystic Ovary Syndrome (PCOS) and Gestational Diabetes Mellitus (GDM) Stratified by Pre-Pregnancy Hypertension Status

Unexpectedly, the results of the stratified analysis suggest that women with PCOS who did not have comorbid pre-pregnancy hypertension had higher prevalence of GDM than women with PCOS and comorbid pre-pregnancy hypertension. However, the results were not statistically significant using the Wald test in models that adjusted for lifestyle factors, reproductive history, and comorbidities. Women with PCOS who did not have comorbid pre-pregnancy hypertension did have significantly higher risk of GDM in all models compared to women without PCOS (Table 3).

^{Table 3. Relationship between polycystic ovary syndrome and gestational diabetes mellitus, stratified by hypertension status}

	Gestational Diabetes (GDM)
	Unadjusted PR (95% CI)	Model 1	Model 2	Model 3
GDM – Birth Certificate
PCOS and Hypertension	1.27 (0.67-2.38)	1.20 (0.62-2.33)	1.01 (0.52-1.98)	1.08 (0.56-2.11)
PCOS without Hypertension	2.27 (1.70-3.03)	2.41 (1.81-3.22)	1.80 (1.34-2.42)	1.60 (1.16-2.21)
Wald Test for Hypertension	0.02	0.04	0.27	0.41
GDM – Survey
PCOS and Hypertension	1.37 (0.79-2.36)	1.13 (0.61-2.10)	0.96 (0.52-1.81)	1.09 (0.59-2.04)
PCOS without Hypertension	2.11 (1.60-2.78)	2.25 (1.71-2.97)	1.72 (1.30-2.28)	1.64 (1.21-2.22)
Wald Test for Hypertension	0.01	0.01	0.08	0.14
GDM – Birth Certificate or Survey
PCOS and Hypertension	1.31 (0.78-2.21)	1.13 (0.63-2.03)	0.97 (0.53-1.76)	1.08 (0.59-1.97)
PCOS without Hypertension	2.10 (1.61-2.74)	2.27 (1.74-2.96)	1.74 (1.33-2.26)	1.61 (1.22-2.14)
Wald Test for Hypertension	0.00	0.00	0.05	0.11

Discussion

The primary finding of this statewide sample of mothers 2-4 months postpartum indicate that women who reported having been diagnosed with preconception PCOS had significantly higher prevalence of GDM compared to women without PCOS when controlling for confounding variables. Hypertension was not found to be a statistically significant effect moderator. This test was done to contribute to our understanding of the underlying mechanism between PCOS and GDM as well as inform clinical care. However, it’s possible the effect moderator was not significant because of the small number of women with both PCOS and hypertension (n=158).

It is also possible that pregnant women with PCOS who do not have comorbid pre-pregnancy hypertension do not receive the same extent of clinical care as women with comorbid conditions and this contributes to their elevated risk of GDM. Women with chronic hypertension, high blood pressure at or above 140/90 mmHg before getting pregnant, are at higher risk of preeclampsia, which is a leading cause of maternal mortality.^7,13 This elevated risk may cause women with pre-pregnancy hypertension to receive more care than pregnant women without hypertension. For example, women with pre-pregnancy hypertension are recommended to receive preconception health care, discuss medications that are safe to take during pregnancy, monitor blood pressure at home, and make lifestyle modifications to promote a healthy diet and regular physical activity.¹³ These interventions designed to reduce the risk of hypertensive disorders of pregnancy (HDP) may also reduce the risk of GDM.

Interpretation

The findings from this study were consistent with the majority of research including three previous meta-analyses. Yu et al. (2016) reported that women with PCOS showed an elevated prevalence of GDM (RR = 2.78, 95% CI: 2.27–3.40).⁶ Khomami et al. (2019) reported women with PCOS had a higher prevalence of GDM (OR = 2.89, 95% CI: 2.37–3.54)¹⁴ and Boomsma et al. (2006). found that PCOS in pregnancy significantly increased the risk of GDM (OR = 2.94, 95% CI: 1.70–5.08).¹⁵ Though the risk of GDM was elevated among women with pre-pregnancy PCOS, the estimates for this study using a population-based cohort of women in Utah were lower than the risk ratios and odds ratios of the meta-analyses by Yu et al. (2016), Khomami et al. (2019), and Boomsma et al. (2006).  This could be due to Utah’s population being a younger and healthier population compared to the populations of the other studies, which were mainly comprised of women recruited from clinical settings.

Limitations

Limitations of this study include the retrospective nature of the PRAMs questionnaire, which limits the ability to infer causality. The questionnaire collects information from postpartum mothers about experiences they had before and during pregnancy which depend on participant’s accurate recall. Social desirability may also have impacted the collection of data on potential confounders since new mothers may not wish to report certain lifestyle behaviors. Another limitation is that that the PRAMS survey is restricted to women who successfully had a live birth. Women with PCOS may be underrepresented because of subfertility. This is important because the results may differ if women who have not yet conceived were included. Classifying women with PCOS based on self-reported diagnosis likely underestimates the number of women with PCOS, compared to using self-reported symptoms to identify women with PCOS. However, this misclassification of the exposure biases the results toward the null. Even more broadly, the lack of a biomarker that can be used in PCOS diagnosis is a significant limitation that affects all research on PCOS. Finally, this dataset did not have information on PCOS phenotype which would have been useful in the analysis.

Strengths

Strengths of this study include the population-based sample which is representative of Utah due to its systematic stratified sampling scheme. This increases the generalizability of the results. The PRAMS questionnaire and linked birth records allowed for detailed information about sociodemographic characteristics, lifestyle, reproductive history, and comorbidities to be used to assess potential confounding. Finally, GDM was measured through the birth certificate, PRAMS survey, and through a combined measure.

Health Implications

These findings indicate that women with PCOS are at elevated risk for GDM, among a population-based sample of mothers in Utah. This is the first known study to test the association between PCOS and GDM in Utah but is consistent with prior literature in other populations indicating women with PCOS are at elevated risk for GDM when controlling for confounding factors. This finding should inform the development of clinical care guidelines, health promotion programs, and policies in Utah and locales with comparable populations. Interventions to reduce the risk of GDM among women with PCOS need to be developed and evaluated.

This is also the first known study to test hypertension as an effect moderator between PCOS and GDM. The purpose of testing hypertension as an effect moderator was to determine if the risk of GDM differs among women with PCOS who do and do not have pre-pregnancy hypertension. In this study, women with PCOS who did not have comorbid pre-pregnancy hypertension appeared to have higher prevalence of GDM than women with PCOS and comorbid hypertension. However, the results were not statistically significant in models that adjusted for lifestyle factors, reproductive history, and comorbidities. Future studies should explore whether in practice women with pre-pregnancy hypertension receive more clinical care than those without pre-pregnancy hypertension. If this is the case, and future research indicates that these interventions lower their risk of GDM, these interventions may also be beneficial to women with PCOS. More research is needed on the role of comorbidities such as hypertension as effect modifiers between PCOS and GDM, the causal mechanisms between PCOS and GDM, and interventions to prevent GDM among women with PCOS.

Acknowledgements

Data were provided by the Utah Pregnancy Risk Assessment Monitoring System, a project of the Utah Department of Health (UDOH), the Office of Vital Records and Health Statistics of the UDOH, and the Centers for Disease Control and Prevention (CDC) of the United States Department of Health and Human Services. The authors extend appreciation to the participants who made the study possible. This report does not represent the official views of the UDOH or CDC.

Funding

Dr. Karen Schliep was supported by a NIH National Institute of Aging (NIA) grant, “Hypertensive Disorders of Pregnancy and Subsequent Risk of Vascular Dementia, Alzheimer’s Disease, or Related Dementia: A Retrospective Cohort Study Taking into Account Mid-Life Mediating Factors” (K01AG058781).

References

1. Azziz, Ricardo. “Introduction: Determinants of Polycystic Ovary Syndrome.” Fertility and Sterility, vol. 106, no. 1, 2016, pp. 4–5, https://doi.org/10.1016/j.fertnstert.2016.05.009.
2. The Rotterdam ESHRE/ASRM-sponsored PCOS consensus workshop group. “Revised 2003 Consensus on Diagnostic Criteria and Long-Term Health Risks Related to Polycystic Ovary Syndrome (PCOS).” Human Reproduction, vol. 19, no. 1, Jan. 2004, pp. 41–47, https://doi.org/10.1093/humrep/deh098.
3. Carson, Sandra Ann, and Amanda N. Kallen. “Diagnosis and Management of Infertility: A Review.” JAMA, vol. 326, no. 1, July 2021, p. 65, https://doi.org/10.1001/jama.2021.4788.
4. Shrivastava, Sneha, and Rosemarie L. Conigliaro. “Polycystic Ovarian Syndrome.” Medical Clinics of North America, vol. 107, no. 2, 2023, pp. 227–34, https://doi.org/10.1016/j.mcna.2022.10.004.
5. Cooney, Laura G., and Anuja Dokras. “Beyond Fertility: Polycystic Ovary Syndrome and Long-Term Health.” Fertility and Sterility, vol. 110, no. 5, 2018, pp. 794–809, https://doi.org/10.1016/j.fertnstert.2018.08.021.
6. Yu, Hai-Feng, et al. “Association between Polycystic Ovary Syndrome and the Risk of Pregnancy Complications: A PRISMA-Compliant Systematic Review and Meta-Analysis.” Medicine, vol. 95, no. 51, 2016, p. e4863, https://doi.org/10.1097/MD.0000000000004863.
7. Mirza, Fadi G., et al. “Polycystic Ovarian Syndrome (PCOS): Does the Challenge End at Conception?” International Journal of Environmental Research and Public Health, vol. 19, no. 22, Nov. 2022, p. 14914, https://doi.org/10.3390/ijerph192214914.
8. Centers for Disease Control and Prevention. PRAMS Methodology. 28 Mar. 2023, https://www.cdc.gov/prams/methodology.htm.
9. Utah Department of Health & Human Services. “Maternal and Infant Health Program.” Utah PRAMS, https://mihp.utah.gov/pregnancy-and-risk-assessment. Accessed 18 July 2023.
10. Shulman, Holly B., et al. “The Pregnancy Risk Assessment Monitoring System (PRAMS): Overview of Design and Methodology.” American Journal of Public Health, vol. 108, no. 10, 2018, pp. 1305–13, https://doi.org/10.2105/AJPH.2018.304563.
11. Ziogas, Christina, et al. “Validation of Birth Certificate and Maternal Recall of Events in Labor and Delivery with Medical Records in the Iowa Health in Pregnancy Study.” BMC Pregnancy and Childbirth, vol. 22, no. 1, 2022, p. 232, https://doi.org/10.1186/s12884-022-04581-7.
12. Hosler, Akiko S., et al. “Agreement Between Self-Report and Birth Certificate for Gestational Diabetes Mellitus: New York State PRAMS.” Maternal and Child Health Journal, vol. 14, no. 5, 2010, pp. 786–89, https://doi.org/10.1007/s10995-009-0529-3.
13. CDC. “High Blood Pressure During Pregnancy.” Centers for Disease Control and Prevention, 19 June 2023, https://www.cdc.gov/bloodpressure/pregnancy.htm.
14. Bahri Khomami, Mahnaz, et al. “Increased Maternal Pregnancy Complications in Polycystic Ovary Syndrome Appear to Be Independent of Obesity—A Systematic Review, Meta‐analysis, and Meta‐regression.” Obesity Reviews, vol. 20, no. 5, 2019, pp. 659–74, https://doi.org/10.1111/obr.12829.
15. Boomsma, C. M., et al. “A Meta-Analysis of Pregnancy Outcomes in Women with Polycystic Ovary Syndrome.” Human Reproduction Update, vol. 12, no. 6, Aug. 2006, pp. 673–83, https://doi.org/10.1093/humupd/dml036.

Citation

Myrer RS, Adediran E, Ellsworth AD, Ceballos RM, Lopez I, Stanford JB, Talboys S, Wang J, Schliep KC. (2024). The association between preconception polycystic ovary syndrome and gestational diabetes mellitus among women with and without pre-pregnancy hypertension: a cross-sectional study from Utah’s Pregnancy Risk Assessment Monitoring System Survey (2016-2021). Utah Women’s Health Review. doi: 10.26054/d-k952-0keb

PDF

View/download

Abstract

Background: Physical inactivity and unhealthy eating are associated with morbidity and mortality, elevating health disparities in underserved communities. This study aims to identify common barriers to fruit/vegetable consumption and physical activity and explore factors associated with these barriers in women of color living in urban Utah.

Methods: The study involves secondary analysis of baseline data from a randomized trial of a wellness-coaching intervention led by Community Health Workers. Study participants from five diverse racial/ethnic communities were interviewed and self-reported demographics, health knowledge and behaviors, and perceived benefits and barriers to healthy behaviors. Self-reported barriers to fruit/vegetable consumption and physical activity were categorized as present or absent. The relationship between number of barriers and demographic factors were analyzed using ANOVA.

Results: Among 485 women in the study, the most reported barriers to fruits/vegetables included quick spoilage (n=203, 41.9%), high cost (n=161, 33.2%), and fruits/vegetables not being a typical part of diet (n=156, 32.2%). The most reported barriers to physical activity included being too tired (n=243, 50.1%), lack of time (n=184, 37.9%), and safety at night (174, 35.9%). More barriers were reported to physical activity than healthy eating. Statistically significant differences in reported barriers were seen between women from different racial/ethnic groups, as well as by age and education.

Conclusions: Among women of color participating in health-behavior change intervention, barriers to fruit/vegetable consumption and physical activity were common and aligned with the seven domains of health. Future interventions should consider these common barriers to improve health behaviors and address health inequity.

_{This work was supported by the Office on Women’s Health, Department of Heath and Human Services under Grant number: 1CCEWH111018-01-00; and National Center for Advancing Translational Sciences, National Institutes of Health under Grant number: UL1TR000105.}

Background

Women’s health is intricately linked to lifestyle factors, particularly physical activity (PA) and nutrition. Lifestyle medicine—an approach that emphasizes healthy lifestyle choices to prevent and treat disease—plays a crucial role in addressing the unique health needs of women. ¹ Healthy eating and PA patterns vary across racial and gender groups, with women of color engaging in reportedly low levels of both.^2-4  Low levels of PA and fruit and vegetable consumption are associated with increased morbidity and mortality, exacerbating health disparities in often underserved communities.^2,5 The Centers for Disease Control and Prevention recommends the consumption of 5-9 servings of fruits and vegetables daily and at least 150 minutes of PA per week to reduce the risk of chronic disease.^6,7  Perceived barriers to healthy living often lead to reduced engagement in these behaviors.^8,9 For example, common reported barriers to fruit and vegetable consumption and PA among African American and Hispanic/Latina women include fatigue, cost, competing family demands, cultural norms, and access among others.^10-14 No research has been reported examining barriers among Pacific Islander/Native Hawaiian, American Indian/Alaskan Native, and African Immigrant/Refugee women, particularly those living in urban areas. Gathering insights in collaboration with women from diverse community groups may elevate the collective understanding of culturally specific barriers. To address this gap in the literature, our study aims to identify common barriers to fruit and vegetable consumption and physical activity among African immigrants and refugees, African Americans, Hispanic/Latino, Pacific Islander/Native Hawaiian, and American Indian/Alaska Native women. Specifically, we will explore the factors associated with these barriers to healthy lifestyle behaviors.

Data

Methodology

The data reported here were gathered as part of a 5-year study was led by the Coalition for a Healthier Community for Utah Women and Girls (UWAG) and community partners including Utah Women’s Health Coalition and Community Faces of Utah, Best of Africa, Calvary Baptist Church, Hispanic Health Care Task Force, National Tongan American Society, Urban Indian Center, University of Utah, and Utah Department of Health. Community Health Workers, referred to as Community Wellness Coaches for this project, recruited participants and collected the data. Each community organization involved had one to two Community Wellness Coaches dedicated to this effort. Participants (n = 485) were asked verbally during in-person computer-assisted baseline interviews (conducted August 31, 2012 through April 17, 2015) about demographics, health knowledge and behaviors, and perceived benefits and barriers to healthy behaviors collected August 31, 2012 through April 17, 2015.

Self-reported barriers collected during interviews were categorized into four groups: nutrition, physical activity, nutrition access, and physical activity access. A six-point response system was used to rate each barrier, with answer choices of strongly agree, somewhat agree, neither agree nor disagree, somewhat disagree, strongly disagree, or not appliable. For data analyses, barriers were classified as present or not present. For positively worded barriers (e.g., I feel safe walking in my neighborhood at night), responses of somewhat disagree or strongly disagree were coded as endorsing the barrier. For negatively worded barriers (e.g., Fruits and vegetables often spoil before I get the chance to eat them), responses of somewhat agree or strongly agree were coded as endorsing the barrier. The barriers used during interviews with participants were based on previous studies and further validated with input from community leaders. The frequency that each barrier was reported is summarized in Figure 1 and Figure 2. The difference in the average number of barriers by category within groups was evaluated by single-factor ANOVA for specific socio-demographic factors including racial/ethnic group, age, education, employment status, federal poverty level (FPL), marital status, and household composition. Statistical analyses were performed using R Studio 3.6.1¹⁵ and Microsoft Excel for Microsoft 365.

Results

The study population included 485 women (ages 18 to 80) from African Immigrant/Refugee (n = 83), Hispanic/Latino (n = 138), Native Hawaiian/Pacific Islander (n = 87), and American Indian/Alaska Native (n = 74) communities who resided in Salt Lake and Utah counties. The study was powered to evaluate results across all communities combined but was not powered to assess associations within each individual community.Therefore, socio-demographic and health factors for all communities of participants are summarized in Table 1.

Sociodemographic Factor	N (%)
Racial/Ethnic Group
African Refugee/Immigrant	83 (17%)
African American	103 (21%)
Hispanic/Latina	138 (28%)
Native Hawaiian/Pacific Islander	87 (18%)
American Indian/Alaska Native	74 (15%)
Age
<25 Years	60 (12%)
25-34 Years	120 (25%)
35-44 Years	131 (27%)
45-54 Years	88 (18%)
55-64 Years	64 (13%)
>65 Years	22 (5%)
Education
Less than High School Degree	74 (15%)
High School Graduate	129 (27%)
Technical School Graduate, Some Technical School, or Certificate	56 (12%)
Associate’s Degree or Some College	128 (26%)
College Graduate or Advanced Graduate Degree	97 (20%)
Employment Status
Full Time	173 (36%)
Part Time or Self-Employed	109 (22%)
Student	29 (6%)
Homemaker	82 (17%)
Retired or Disabled	39 (8%)
Unemployed or Looking for Work	83 (17%)
Federal Poverty Level (FPL)
100% or Less of the FPL	224 (46%)
101% – 130% of the FPL	55 (11%)
131% – 185% of the FPL	71 (15%)
>185% of the FPL	102 (21%)
Not Reported	33 (7%)
Marital Status
Single	140 (29%)
Married	220 (45%)
Living with Partner, Not Married	43 (9%)
Divorced, Separated, or Widowed	81 (17%)
Children in the Home	140 (29%)
No Children <18	162 (33%)
1 Child <18	83 (17%)
2 Children <18	97 (20%)
3 Children <18	68 (14%)
≥4 Children <18	74 (15%)
Elders in the Home
No Elders 65+	434 (90%)
Any Elders 65+	50 (10%)
Health Insurance Access
Medicaid	112 (23%)
Other	186 (38%)
Uninsured	185 (38%)
Body Mass Index
Underweight/Normal Weight	81 (17%)
Overweight	127 (26%)
Obese	276 (57%)
Self-Rated Health Status
Excellent/Very Good/Good	255 (53%)
Fair/Poor	223 (46%)
Depression
Positive PHQ-2 Depression Screen (≥3)	105 (22%)
Number of Reported Barriers at Baseline	(Mean (SD), Range)
Barriers to Healthy Eating (0-12)	2.08 (1.95), 0-9
Barriers to Accessing Healthy Foods (0-5)	0.59 (0.98), 0-4
Barriers to Physical Activity (0-14)	3.38 (2.84), 0-14
Barriers to Physical Activity Access (0-8)	1.42 (1.44), 0-7
Number that Reported Zero Barriers
Barriers to Healthy Eating	115 (24%)
Barriers to Accessing Healthy Foods	321 (66%)
Barriers to Physical Activity	101 (21%)
Barriers to Physical Activity Access	161 (33%)

The most commonly reported barriers to healthy eating and healthy eating access were quick spoilage of healthy food (n = 203, 41.9%), high cost of healthy food (n = 161, 33.2%), healthy foods not a typical aspect of diet (n = 156, 32.2%), healthy foods not filling enough (n = 121, 24.9%), and family dislikes healthy foods (n = 81, 16.7%). A full summary of reported healthy eating and healthy eating access barriers are shown in Figure 1.

The most commonly reported barriers to physical activity and physical activity access were being too tired to be physically active (n = 243, 50.1%), not having time to be physically active (n = 184, 37.9%), not feeling safe walking in neighborhood at night (n = 174, 35.9%), too many family demands (n = 162, 33.4%), and no partner to exercise with (n = 149, 30.7%). A larger percentage of participants reported PA access barriers in comparison to healthy eating and healthy eating access barriers. A full summary of reported physical activity and physical activity access barriers are shown in Figure 2.

Comparing the average number of barriers reported by category between groups, there was a statistically significant difference in average number of barriers across racial/ethnic groups for nutrition barriers (p <0.00001), nutrition access barriers (p = 0.00003), and physical activity access barriers (p <0.00001). Women in Hispanic/Latino communities reported the highest average number of nutrition barriers (2.66), whereas African immigrant/refugee women reported the lowest average (1.18). Similarly, Hispanic/Latino women faced the greatest number of nutrition access barriers (0.82), while Native Hawaiian/Pacific Islander women reported the fewest (0.33). In terms of physical activity access barriers, women in American Indian/Alaskan Native communities reported the highest average (2.03), in contrast to Native Hawaiian/Pacific Islander women, who reported the lowest average (0.94).

Additionally, age groups exhibited significant differences in the average number of physical activity barriers (p = 0.00683). Specifically, 35–44-year-olds reported the highest average (4.05), while those over 65 years reported the lowest (2.50). Educational attainment also correlated with nutrition access barriers, with significant differences observed (p = 0.00395). technical school graduates, those with some technical school training, or a certificate training reported the highest average number of nutrition access barriers (3.05), whereas individuals with associate degrees or some college reported the lowest average (0.43). There were no significant differences in the average number of reported barriers based on employment status (Full Time, Part Time/Self Employed, Student, Homemaker, Retired/Disabled, Unemployed/Looking for Work), federal poverty level, marital status (Single, Married, Living with Partner/Not Married, Divorced/Separated/Widowed), or household composition (No Children <18, Children <18, No Elders 65+, Elders 65+). Further information on the average number of barriers reported by demographic groups is shown in Table 2.

Demographics	Nutrition Barriers (Mean (SD))	Nutrition Access Barriers (Mean (SD))	Physical Activity Barriers (Mean (SD))	Physical Activity Access Barriers (Mean (SD))
Ethnic Group
African	1.18 (1.32)	0.82 (1.24)	2.87 (2.73)	1.16 (1.21)
African American	1.73 (1.74)	0.36 (0.85)	3.06 (2.55)	1.26 (1.42)
Hispanic/Latino	2.66 (2.17)	0.83 (1.04)	3.83 (3.00)	1.70 (1.55)
Native Hawaiian/Pacific Islander	2.39 (1.90)	0.33 (0.71)	3.48 (2.95)	0.94 (1.18)
American Indian/Alaskan Native	2.18 (2.06)	0.49 (0.80)	3.70 (2.85)	2.03 (1.51)
p-Value	< .00001	0.000028	0.074553	< .00001
Age
<25 Years	1.73 (1.61)	0.68 (1.10)	2.57 (2.59)	1.35 (1.39)
25-34 Years	2.32(2.07)	0.50 (0.93)	3.58 (2.98)	1.37 (1.36)
35-44 Years	2.39 (1.90)	0.79 (1.06)	4.05 (2.89)	1.49 (1.51)
45-54 Years	2.15 (2.17)	0.49 (0.84)	3.39 (2.79)	1.52 (1.48)
55-64 Years	1.78 (1.90)	0.42 (0.91)	3.00 (2.62)	1.25 (1.38)
>65 Years	1.73(1.98)	0.45 (0.91)	2.50 (2.54)	1.73 (1.70)
p-Value	0.126124756	0.060286926	0.00683965	0.713430489
Education
Less than High School	2.15 (2.16)	0.62 (1.03)	3.31 (3.06)	1.38 (1.32)
High School Graduate	2.40 (2.00)	0.51 (0.88)	3.40 (2.95)	1.30 (1.34)
Technical School Graduate, Some Technical School, or Certificate	1.95 (1.83)	1.02 (1.23)	3.05 (2.91)	1.48 (1.43)
Associate’s Degree or Some College	2.01 (1.91)	0.43 (0.85)	3.73 (2.86)	1.67 (1.60)
College Graduate or Advanced Graduate Degree	1.82 (1.84)	0.63 (0.99)	3.19 (2.40)	1.28 (1.44)
p-Value	0.223731	0.003955	0.519894	0.21151

Barriers to a Healthy Lifestyle and the 7 Domains

The data reported here align directly with six of the seven domains of health.^16,17 Emotional health barriers were reported, including body image concerns (“don’t look good while exercising”) and lack of prioritizing healthy eating (“not as important as other things”). Environmental health barriers were highlighted for both nutritional factors (“spoil/waste” and “stores in neighborhood don’t sell [fruits and vegetables]”) and physical activity factors (“too much pollution or noise” and “weather makes it hard to exercise”). Financial barriers (high cost or too expensive) were highly reported for nutritional behaviors, with less individuals reporting cost as a barrier to physical activity. Intellectual health barriers included lack of skill in both choosing and preparing fruits and vegetables as well as knowing what to do during exercise. Physical health barriers to health eating behaviors included disliking fruits and vegetables, having a hard time chewing or digesting them, and a lack of knowledge about the benefits associated with consuming fruits and vegetables. The barriers identified by the women were not associated directly with the spiritual health domain. However, in an open-response question, individuals expressed barriers due to a lack of motivation. Exploring motivational factors associated with spiritual health may be valuable to explore in future research.

Resources and Future Recommendations

Examining the barriers associated with physical activity and healthy eating behaviors among women of color provides insights for future research, community programming, and healthcare services. Efforts to reduce the perceived barriers among this unique population have the potential to encourage healthy lifestyle behaviors, reduce the risk and prevalence of chronic diseases, and reduce health inequities in Utah communities.

Several key strengths identified in this study provide guidance for future research and programming. The study examined barriers across various cultural groups, acknowledging that distinct groups of women of color perceive unique barriers to healthy eating and activity. Working as a community-based team that included community leaders, Community Wellness Coaches, and researchers was a key strength of this study. Engaging community members in the research process ensures that the findings are relevant and applicable to the populations being studied. This collaboration fosters trust, enhances the quality of data collected, and empowers community voices, making future research more impactful. Additionally, the analysis performed in this study, which included ANOVA to compare the average number of barriers reported across different demographic groups and the calculation of percentages of participants who reported each barrier, adds rigor to the findings. This multifaceted approach not only highlights disparities but also provides a comprehensive understanding of the barriers faced by diverse groups.

Despite the strengths, several limitations were also identified. Combining vegetable and fruit consumption in the survey limited our ability to examine the dietary habits across the two types of food. Reported perceived barriers to consumption, for example, may differ across fruits and vegetables.  Findings from the present study also highlight the diverse interpretations of what the term ‘health’ and ‘physical activity’ mean for each community. While the study provides a valuable For example, physical activity can mean different things in different communities; it might refer to organized sports in one context and daily chores or informal play in another. Analyzing interview responses highlights the need to define ‘physical activity’ more inclusively—covering exercise, daily tasks, and general movement. This would improve transparency, standardize question interpretation, and reduce assimilation bias.examination of perspectives from five unique communities, future research exploring the community story qualitatively is recommended. Finally, for migrant communities, barriers related to access to resources may differ between individuals who recently arrived and those who have been resettled for a longer period. We propose a number of programming strategies in response to awareness of perceived barriers. To effectively support the nutritional needs of women of color, research shows it is essential to create culturally relevant nutritional education initiatives, evaluate food assistance programs, and implement culturally appropriate screening in clinical settings for each community.⁹  These types of programs can help reduce food insecurity and disparities. Specific to physical activity, designing programs that address broader patterns of time use, including improving sleep time and reducing sedentary behavior time, may mediate the relationship between physical activity and health for women of color.¹⁸ Again, physical activity programs need to be culturally tailored. Additional research suggests tailoring enrollment efforts to at-risk women who do not meet the physical activity and/or dietary guidelines to improve cost effectiveness¹⁹  and acknowledging the commonalities and differences across culture and gender related to lifestyle behaviors.²⁰ As a follow-up to this study, we propose conducting community-based participatory qualitative research with community leaders to explore how the findings align with or differ from community lived experiences.

References

1. Geyer C, McHugh J, Tollefson M. Lifestyle Medicine for Women: The Time Is Now! American Journal of Lifestyle Medicine. 2021;15(4):366-371. doi:https://doi.org/10.1177/15598276211004233.
2. Bennett G, Bardon LA, Gibney ER. A Comparison of Dietary Patterns and Factors Influencing Food Choice among Ethnic Groups Living in One Locality: A Systematic Review. Nutrients. 2022;14(5):941. doi:https://doi.org/10.3390/nu14050941.
3. Blackwell D, Villarroel M. Tables of Summary Health Statistics for US Adults: 2017 National Health Interview Survey. National Center for Health Statistics. 2018.
4. Lorenzo E, Szeszulski J, Todd M, Mama SK, Lee RE. Health Is Power: Active transportation, physical activity, and cardiometabolic health among ethnic minority women. Journal of physical activity and health. 2020;17(3):323-330. doi:https://doi.org/10.1123/jpah.2019-0098.
5. 2018 Physical Activity Guidelines Advisory Committee Scientific Report. (U.S. Department of Health and Human Services) (2018).
6. Lee SH, Moore LV, Park S, Harris DM, Blanck HM. Adults Meeting Fruit and Vegetable Intake Recommendations — United States, 2019. MMWR Morb Mortal Wkly Rep 2022;71:1–9. DOI: http://dx.doi.org/10.15585/mmwr.mm7101a1.
7. Hall KD, Sacks G, Chandramohan D, et al. Quantification of the effect of energy imbalance on bodyweight. The Lancet. 2011;378(9793):826-837. doi:https://doi.org/10.1016/S0140-6736(11)60812-X
8. Trilk J, Nelson L, Briggs A, Muscato D. Including lifestyle medicine in medical education: rationale for American College of Preventive Medicine/American Medical Association resolution 959. American Journal of Preventive Medicine. 2019;56(5):e169-e175. doi:https://doi.org/10.1016/j.amepre.2018.10.034.
9. Willis SK, Simonsen SE, Hemmert RB, Baayd J, Digre KB, Zick CD. Food Insecurity and the Risk of Obesity, Depression, and Self-Rated Health in Women. Womens Health Rep (New Rochelle). 2020;1(1):308-317. Published 2020 Aug 31. doi:10.1089/whr.2020.0049.
10. Bertoni AG, Foy CG, Hunter JC, Quandt SA, Vitolins MZ, Whitt-Glover MC. A multilevel assessment of barriers to adoption of Dietary Approaches to Stop Hypertension (DASH) among African Americans of low socioeconomic status. Journal of health care for the poor and underserved. 2011;22(4):1205-1220. doi:10.1353/hpu.2011.0142.
11. Gordon-Larsen P, Nelson MC, Page P, Popkin BM. Inequality in the built environment underlies key health disparities in physical activity and obesity. Pediatrics. 2006;117(2):417-424. doi:https://doi.org/10.1542/peds.2005-0058.
12. Lucan SC, Barg FK, Long JA. Promoters and barriers to fruit, vegetable, and fast-food consumption among urban, low-income African Americans—a qualitative approach. American journal of public health. 2010;100(4):631-635. doi:https://doi.org/10.2105/AJPH.2009.172692.
13. Shuval K, Leonard T, Murdoch J, Caughy MO, Kohl HW, Skinner CS. Sedentary behaviors and obesity in a low-income, ethnic-minority population. Journal of Physical Activity and Health. 2012;10(1):134-138. doi:https://doi.org/10.1123/jpah.10.1.134.
14. Zenk SN, Odoms-Young AM, Dallas C, et al. “You have to hunt for the fruits, the vegetables”: environmental barriers and adaptive strategies to acquire food in a low-income African American neighborhood. Health education & behavior. 2011;38(3):282-292. doi:https://doi.org/10.1177/1090198110372877.
15. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2021. https://www.R-project.org/.
16. Frost, C.J. & Digre, K.B., eds. (2016). The 7 Domains of Health:  Multidisciplinary Considerations of Women’s Health in the 21^st Century. Dubuque, Iowa:  Kendall Hunt Publishers. 
17. Frost, C., Murphy, P., Shaw, J., Jones, K., Varner, M….Digre, K. (2013). Reframing the view of women’s health in the United States: Ideas from a multidisciplinary national center of excellence in women’s health demonstration project. Clin Mother and Child Health, 11(01). https://doi.org/10.4172/2090-7214.1000156.
18. Zick CD, Buder I, Waitzman NJ, Simonsen S, Digre K. The nexus between health and time use among racially and ethnically diverse women. Ethnicity & Health. 2019;24(2):147-167. doi:https://doi.org/10.1080/13557858.2017.1315529.
19. Buder I, Zick C, Waitzman N, Simonsen S, Sunada G, Digre K. It takes a village coach: Cost-effectiveness of an intervention to improve diet and physical activity among minority women. Journal of Physical Activity and Health. 2018;15(11):819-826. doi:https://doi.org/10.1123/jpah.2017-0285
20. Simonsen SE, Digre KB, Ralls B, et al. A gender-based approach to developing a healthy lifestyle and healthy weight intervention for diverse Utah women. Evaluation and Program Planning. 2015;51:8-16. doi:https://doi.org/10.1016/j.evalprogplan.2014.12.003.

Citation

Fernandez M, O’Farrell KD, Digre K, Stark L, Davis FA, Lee D, Mukundente V, Napia E, Sanchez-Birkhead A, Tavake-Pasi F, Villalta J, Brown H, & Simonsen SE. (2024). Barriers to Healthy Eating and Physical Activity for Women of Color in Utah. Utah Women’s Health Review. doi: 10.26054/d-6ffd-2f44

PDF

View/download