Association of types of diabetes and insulin dependency on birth outcomes

lNTRODUCTlON

Diabetes mellitus(DM)rates in the United States have been increasing and women with diabetes in pregnancy have high rates of congenital anomalies,preeclampsia,preterm delivery,macrosomia,and perinatal mortality[1-7].In the United States,approximately seven percent of pregnancies are affected by DM,a condition in which a woman’s blood glucose levels are above normal.Traditionally,DM has been classified into one of three categories:Type 1 diabetes,type 2 diabetes,and gestational diabetes(GDM)[8].All three forms result in the body’s inability to produce enough insulin,the hormone responsible for cells taking in sugar from the bloodstream to be stored and later used as energy.Type 1 diabetes accounts for about 5% of all diabetes cases,mainly caused by the autoimmune system’s attack on beta cells that create insulin-dependence,the etiology of which involves both genetic and environmental risk factors.Type 2 diabetes is not an autoimmune condition,but rather a metabolic condition often related to obesity,a sedentary lifestyle and poor diet,where the body loses its ability to respond to insulin,creating insulin-resistance[9,10].GDM is diabetes that is diagnosed in a woman during pregnancy and accounts for about 85% of diabetes cases among pregnant women,also associated with genetic and environmental factors,with incidence rates also rising[11-13].As the prevalence of DM is increasing among women,so too grows the public health threat this health condition poses to pregnancy and birth outcomes[14,15].

The effect of insulin resistance on birth outcomes has been well documented.For example,Klemetti[16]analyzed hospital data from 881 pregnant patients with type 1 diabetes,over a ten-year span(1998-2008),and found that poor glycemic control was associated with increases in emergency caesarean sections,macrosomia,and Neonatal Intensive Care Unit(NICU)admission rates.One large population study in Denmark found that among women with type 1 diabetes,those with the greatest glycemic control before pregnancy had the lowest risk for poor birth outcomes,including perinatal mortality and serious adverse outcomes[17].Others report that when evaluating type 2 diabetes,those with preconception care had lower rates of fetal malformations than those without preconception care[18].A large meta-analysis supports the claim that there is evidence of increased pre-eclampsia,cesarean delivery,and macrosomia for women with type 1 diabetes that have poor glycemic control[19].Conversely,strong glucose control among type 1 diabetes has been associated with decreased risk for perinatal mortality,decreased maternal hypoglycemia,and normal fetal weight[20-22].

The backdoor opened. My husband walked outside. I love you, he said. I was wondering where you were. I chuckled11, as he came over and kissed me, then turned around and went back inside. Where s the quiet time? I asked God. My heart fluttered. There was no pain, only a beat that interrupted me yet again. This is impossible, I thought. There s no time to be still and to know that God is with me. There s too much going on in the world and entirely12 too much activity all around me.

There is an important gap in the published literature,however,regarding population level studies of diabetes during pregnancy.For example,several large cohort studies of birth certificate data have reported an association between diabetes and birth outcomes,although differences in birth outcomes between diabetes types have rarely been reported[5,15,23,24].Further,when types of DM have been compared,the results have been inconsistent[6].For example,one study comparing type 1 diabetes to type 2 diabetes reported pregnant women with type 1 diabetes had an increased risk for preterm birth,large for gestational age,and hypertension[25].Others have reported no differences between pregnant women with type 1 diabetes and type 2 diabetes as it relates to rates of congenital malformations and perinatal mortality[26,27].Still another study found that still births and congenital anomalies were highest in the type 1 diabetes group with the lowest glycemic levels,although women with GDM were not included in that study[28].One population-based study in France reported that preterm birth and macrosomia rates were significantly higher for women with type 1 diabetes(ORs 5.8 and 7.7,respectively),type 2 diabetes(ORs 3.1 and 3.8,respectively),and GDM(ORs 1.2 and 1.8,respectively).Further,in that study,there was 95% confidence that the odds of perinatal mortality for women with type 1 diabetes were comparable with non-diabetic women(OR 1.8,95%CI:1.0,3.1),yet significantly higher for women with type 2 diabetes(OR 3.6,95%CI:2.6,5.0)and significantly lower yet for women with GDM(OR 0.70,95%CI:0.60,0.80)[29].A major limitation to that population based study in France was that they did not have any data on body mass index,an important variable related to diabetes and strong potential confounder.Given the inconsistent results of these previous studies,additional research is needed to clarify the differential impact of each type of DM on birth outcomes.

Xaverius PK oversaw all aspects of this project including developing the research question,collecting and analyzing the data,and writing the overall manuscript;Xaverius PK and Kiel D designed the research study;Wankum E,Carter C,Fang C,and Carriere R analyzed the data;and Xaverius PK,Howard SW,and Thurman JE wrote the manuscript;All authors have read and approve the final manuscript.

Diabetes is associated with the poor birth outcomes.Clinical management of diabetes during pregnancy and healthy lifestyle behaviors before pregnancy can reduce the risk for diabetes and poor birth outcomes.

MATERlALS AND METHODS

Study design

We conducted a population-based cross-sectional study of live,singleton births in Missouri from 2010-2012 inked with death certificate data.We removed implausible BMI categories(＜ 12 and ＞ 70 BMI),resident zip codes outside the state of Missouri,and gestational age less than 20 wk,bringing our sample to 207511[32].Cases were also removed when birth weight,race/ethnicity,marital status,smoking status,maternal education,and maternal age were missing(1.7%),resulting in a final sample of 204057.Listwise deletion was used,as there was sufficient sample size to support removing data that was missing at random and the percent of cases that were removed was less than 5% of the overall sample[33,34].

Exposure

Diabetes categories were selected based upon diabetes status and insulin-dependence status identified on the birth certificate.Prepregnancy diabetes non-insulin dependent(PD-NI),or insulin dependent(PD-I),gestational diabetes-insulin dependent(GD-I)and non-insulin dependent(GD-NI)and nondiabetic.Birth certificate recorders gather DM information from prenatal records,and this data has been reported elsewhere as having moderate sensitivity[35].

Now, the dove was nearly driven distracted at the jackal s words; but, in order to save the lives of the other two, she did at last throw the little one out of the nest

Covariates

Demographic characteristics included:maternal age(＜ 19,19-34,＞ 34),race/ethnicity(non-Hispanic white,non-Hispanic black,Hispanic,and Other),education(less than high school,high school or GED,some college,college grad or more),adequacy of prenatal care(inadequate,intermediate,adequate,adequate plus,and unknown,based upon the Kotelchuck index),marital status,and Medicaid status.Maternal risk covariates included BMI category[＜ 18.5(underweight),18.5-24.9(normal),25-29.9(overweight),≥ 30(obese)],smoking during pregnancy,prepregnancy hypertension,gestational hypertension,preeclampsia,and previous preterm birth.Covariates often found to correlate with adverse birth outcomes were selected.

Outcomes

Preterm birth(gestational age ＜ 37 wk),macrosomia(birth weight ＞ 4000 g),and infant mortality(death＜ 365 d of age).

Statistical analysis

Crude and multivariate logistic regressions were calculated to assess differences between diabetes groups in outcomes,with 99.8% confidence intervals calculated to measure precision.Bonferroni correction was used for the twenty-four 95%CIs to give a nominal confidence level of 99.8%.Chi-square tests were used to assess differences in selected covariates.Cramer’s V was used to assess the magnitude of the relationship between categorical variables where 0.1 indicates a weak relationship,0.3 indicating a moderate relationship,and 0.5 indicating a strong relationship.

To investigate differences in birth outcomes(preterm birth,macrosomia,and infant mortality/)by diabetes status.

RESULTS

A larger proportion of non-diabetic women were under age 35,and white,non-Hispanic,when compared to diabetic women.In addition,non-diabetic women reported a higher proportion of women that received adequate prenatal care and lower proportion of women that received adequate-plus prenatal care,in comparison with diabetic women.In contrast,a significantly higher proportion of women with PD-I(51.4%),PD-NI(59.4%),GD-I(64.4%)and GD-NI(46.1%)reported 30+ BMI category(obese)and a lower proportion of normal weight gain during pregnancy(20.2%,21.3%,20.4%,and 25.8%,respectively)in comparison with non-diabetic women(22.5% and 28.2%,respectively)(Table 1).

Table 2 presents the crude and adjusted odds ratios for the relationship of diabetes status to each birth outcome.PD-I,PD-NI,GD-I,and GD-NI were significantly associated with an increased risk for preterm birth[cOR 4.15(95%CI:3.22,5.34);cOR 2.45(95%CI:1.85,3.24);cOR 2.54(95%CI:1.91,3.39);and cOR 1.41(95%CI:1.27,1.56);respectively].PD-I,PD-NI,GD-I,and GD-NI were significantly associated with an increased risk for macrosomia[cOR 3.12(95%CI:2.19,3.77);cOR 2.32(95%CI:1.74,3.10);cOR 2.15(95%CI:1.58,2.92);and cOR 1.57(95%CI 1.42,1.74),respectively].The risk for infant mortality was significantly decreased for women in the GD-NI category[PD-I:cOR 1.87(95%CI:0.66,5.28);PD-NI:cOR 1.90(95%CI:0.71,5.10);GD-I:cOR 1.62(95%CI:0.54,4.88);and GD_NI:cOR 0.51(95%CI:0.29,0.89)],in reference to non-diabetic women.There were significantly different crude odds between both the PD-I and GD-NI categories for preterm birth and macrosomia,in comparison to all other preterm birth categories,while the PD-NI and GD-I categories were non-significantly different from each other.Only women within the GD-NI category had significantly different infant mortality rate in reference to non-diabetic women.

Table 2 also presents the adjusted odds ratios for the relationship of diabetes status to each birth outcome,adjusted for maternal age,race/ethnicity,maternal education,marital status,BMI.In the adjusted model,preterm birth remained significantly associated in three of the four categories:PD-I(2.87,95%CI:2.19,3.77),PD-NI(1.77,95%CI:1.31,2.39),GD-I(1.73,95%CI:1.27,2.35),while there was no significant difference in the GD-NI category(1.07,95%CI:0.96,1.19).Macrosomia also remained significantly associated with all four categories:PD-I(3.01,95%CI:2.26,4.01),PD-NI(2.12,95%CI:1.57,2.86),GD-I(1.96,95%CI:1.42,2.70),and GD-NI(1.53,95%CI:1.39,1.70).Women with GD-NI were found to have a significantly lower risk for infant mortality[aOR 0.41(0.23,0.72)]in reference to nondiabetic women.There were significantly different adjusted odds between both the PD-I and GD-NI categories for preterm birth and macrosomia,in comparison to all other preterm birth categories,while the PD-NI and GD-I categories were non-significantly different from each other.

DlSCUSSlON

This large population-based study found that women with pregestational DM had the highest increased risk for preterm birth and macrosomia,with the highest risk for women with insulin dependent prepregnancy diabetes,in reference to non-diabetic women.Statistically,there was no significant difference between the risks for preterm birth and macrosomia among women with PD-NI and GD-I.Interestingly,women with gestational diabetes that had no insulin-dependence were at a 59% reduced risk for infant mortality in comparison with women without diabetes.These findings are fairly comparable to the large population-based study in France,although our odds are slightly lower perhaps due to our ability to include BMI in our adjusted model[33].With the growing prevalence of DM,the differential impact of type of diabetes on birth outcomes is important to identify so evidence-based plans can be implemented to reduce the deleterious impact of this growing public health crisis.

Interestingly,differences in risks between types of DM seem to reflect differences in timing of DM(pre-pregnancygestational)and insulin use.For example,the odds for preterm birth and macrosomiawere higher when there was insulin-dependence,among women with pre-pregnancy diabetes(PTB:187%77%;Macrosomia:126%57%)and among women with gestational diabetes(PTB:73%9%;Macrosomia:96%53%)in comparison with women without diabetes.The risks were significantly higher when comparing insulin dependence within the gestational categories,because the 99.8%confidence intervals did not overlap between gestational categories,and further,there was no difference in risk for preterm birth between women without diabetes and women with non-insulin dependent gestational diabetes.These results reflect similar findings among other published studies,that is,more severe adverse birth outcomes are associated with worsening glycemic control.While measures of glycemic control are not provided in the birth certificate data used for this study,glycemic control is often harder to achieve in women with absolute insulin resistance,perhaps explaining the difference in

poor birth outcomes among insulin use in comparison with non-insulin dependent women[36].

It is notable that women with GD-NI were found to be at a 59% significantly decreased odds for infant death in comparison with non-diabetic women,even after adjusting for socio-demographic,behavioral,and biological differences.This finding is consistent with a study in France,that found a 30% reduction in perinatal mortality among women with GDM in comparison with non-diabetic women[33].Diagnosis of GDM is usually based upon results from oral glucose tolerance tests often conducted between the 24and 28week of gestation.Treatment is designed to lower glucose concentrations and typically involves high-risk obstetric management,including behavioral changes,nutritional plans,or insulin,as needed[37].In the large population-based study from France discussed earlier,when their sample was limited to full-term deliveries that excluded cases of undiagnosed pregestational diabetes,the odds of perinatal mortality reversed from being decreased to being increased when compared with non-diabetic women.This led the authors to speculate that the timing and delivery of treatment may play a pivotal role in reducing risks for infant mortality.We similarly speculate that timing and intensity of GDM treatment play an important role in infant mortality among our study participants in the United States.

There are a number of limitations to this study,which include the possible misclassification of diabetes.First,women with PD-I or PD-NI may have first been diagnosed during pregnancy,and thus their DM status was wrongly classified as gestational.While extremely unlikely for women with PD-I,given the significant symptoms and typical younger age at onset associated with this condition,it is possible that women with PD-NI were undiagnosed before pregnancy and thus classified as GDM.In addition,it is possible that there are some women in the non-diabetic group who were diabetic.Prediabetes is also a growing population-level concern(women with prediabetes have higher blood glucose levels than normal,but not high enough to be medically diagnosed with diabetes[38]),but that data is not provided on birth certificates.Women with prediabetes are at increased risk for developing both GD and non-insulin dependent diabetes mellitus(PD-NI)later in life;this may distort the risk in the non-diabetic group,as individuals labeled as non-diabetic could be pre-diabetic(with increased risk for adverse birth outcomes).In addition,while the result for infant mortality was not significant in the PD-I,PD-NI,and GD-I categories,the point estimates were higher when compared to non-diabetic women and this may be due to the small number of people in those categories.Furthermore,there is potential for residual confounding within the data due to unmeasured behavioral risk factors,income levels,as well as other unmeasured c,that may impact the overall outcomes for these women and their babies.Also,due to the sample only coming from the state of Missouri,there is limited generalizability.Because the data set includes all Missouri births from 2010-2012,this study has strong internal validity with respect to risks of adverse birth outcomes by category of DM for women in Missouri.

CONCLUSlON

As categories of diabetes differed,so too did risk for poor birth outcomes,with having insulin use among women with pre-pregnancy diabetes putting women at the highest risk for the poorest birth outcomes.Clinical management of DM and healthy lifestyle behaviors before pregnancy have been shown to improve birth outcomes,suggesting that access to preconceptional care plays an important role in reducing risks for poor birth outcomes.Clinical implications from these findings should recognize the increased risk for adverse birth outcomes for all categories of diabetes,especially for preterm birth and macrosomia.The classification schema of insulin-dependent,non-insulin dependent,and gestational diabetes may be outdated[6],yet the risks for poor birth outcomes were significantly increased based upon timing of DM onset(,prepregnancy or gestational)and insulin use.We now have a better understanding of the spectrum of factors associated with different forms of DM,including age,weight,metabolic syndrome,autoimmune disease,inflammation,and c-peptide[39].Future research should focus on maintaining proper glycemic control before pregnancy and throughout pregnancy to help reduce the risk for adverse birth outcomes.Further,findings from a large systematic review found that a diet high in fruits and vegetables,legume,nuts,whole grains,and fish before pregnancy may reduce one’s risk for developing GDM during pregnancy[40].Future research should consider how non-pregnant women of childbearing age are assessed and provided education on management and prevention of DM,specifically as it relates to pregnancy.Increased knowledge and implementation of evidence-based standards of care during the preconceptional period could result in reduced rates of DM among women,and in turn,healthier moms and babies.

ARTlCLE HlGHLlGHTS

Research background

Diabetes mellitus(DM)rates in the United States have been increasing and women with diabetes in pregnancy have high rates of congenital anomalies,preeclampsia,preterm delivery,macrosomia,and perinatal mortality.In the United States,approximately seven percent of pregnancies are affected by DM,a condition in which a woman’s blood glucose levels are above normal.The effect of insulin resistance on birth outcomes has been well documented.There is an important gap in the published literature,however,regarding population level studies of diabetes during pregnancy.For example,several large cohort studies of birth certificate data have reported an association between diabetes and birth outcomes,although differences in birth outcomes between diabetes types have rarely been reported.

When my father came home there was laughter, rollicking（，）, rolling laughter. He was strong and handsome; his thick, black, wavy1 hair fell into his black, laughing eyes. When he kissed me, I pushed his bristled2 mustache from my tender skin. His hands, thick and squared off at the tips, smelled of the sweet horsehair at the upholstery（） factory. His fingernails carried the cotton lint3 he used to stuff stain sofas.

Research motivation

Understanding the differential impact of prepregnancy diabetes with and without insulin dependence and GDM can offer important clues to understanding the population impact of insulin dependence on birth outcomes in the United States.This study explores how birth outcomes vary for women exposed based upon timing of diabetes(pre-gestational or gestational)and insulin-dependence,building upon previous studies by including potentially important confounders like BMI(a reliable measure for population-based surveillance).

Research objectives

The next day, my four children and I landed on the Big Island of Hawaii with less than two thousand dollars, knowing no one in the world was going to help us

Thanks for being so sweet, kind and thoughtful yesterday. I was sincerely touched by you. It is so nice to meet someone that’s genuinely nice, warm and sensitive and unselfish. Please don’t change your ways because I truly believe that you will excel. Have a great day! Hank

Research methods

Cross-sectional design,using linked Missouri birth and death certificates[singleton births only),2010 to 2012(= 204057).Exposure was diabetes(non-diabetic,pre-pregnancy diabetes-insulin dependent(PDI),pre-pregnancy diabetes-non-insulin dependent(PD-NI),gestational diabetes-insulin dependent(GDI),and gestational diabetes-non-insulin dependent(GD-NI)].Outcomes included preterm birth,macrosomia,and neonatal death.Confounders included demographic characteristics,adequacy of prenatal care,BMI,smoking,hypertension,and previous preterm birth.Bivariate and multivariate logistic regression assessed differences in outcomes by diabetes status.

Research results

By the beard of the Prophet I never saw such a thing in all my life! My very humble31 thanks, replied the Grand Vizier, as he bent his long neck; but, if I may venture to say so, your Highness is even handsomer as a stork than as a Caliph

Research conclusions

As categories of diabetes differed,so too did risk for poor birth outcomes,with having insulin use among women with pre-pregnancy diabetes putting women at the highest risk for the poorest birth outcomes.

Research perspectives

Understanding the differential impact of prepregnancy diabetes with and without insulin dependence and GDM can offer important clues to understanding the population impact of insulin dependence on birth outcomes in the United States.The newest United States standard birth certificate allows for the examination of DM as it is related to birth outcomes,based upon timing of DM(prepregnancy DM or GDM)and pre-pregnancy DM insulin dependence(insulin dependent DM:PD-I and non-insulin dependent DM:PD-NI)and GDM insulin-dependence(GD-I)and non-insulindependence(GD-NI).This study will explore how birth outcomes vary for women exposed to different categories of DM,marked PD-I,PD-NI,GD-I,and GD-NI and build upon previous studies by including potentially important confounders like body mass index(BMI)(a reliable measure for population-based surveillance)[30,31].

ACKNOWLEDGEMENTS

This project was completed in the epidemiology capstone course as part of the MPH degree requirements at Saint Louis University,College for Public Health and Social Justice,a course coinstructed by Pamela K Xaverius,PhD,MBA and Joanne Salas,MPH.

It happened that where the Prince s father lay buried, there grew a rosetree--a most beautiful rose tree, which blossomed only once in every fiveyears, and even then bore only one flower, but that was a rose! It smelt2 sosweet that all cares and sorrows were forgotten by him who inhaled3 itsfragrance.

FOOTNOTES

Taram-taq was left alone; he would have retired130 into his fort, but the prince shouted: Whither away, accursed one? Are you fleeing before me? At these defiant131 words the chief shouted back, Welcome, man! Come here and I will soften132 you to wax beneath my club

This study has been granted an exemption by the Institutional Review Board at Saint Louis University.

Women with PD-I,PD-NI,and GD-I remained at a significantly increased odds for preterm birth(aOR 2.87;aOR 1.77;and aOR 1.73,respectively)and having a very large baby(macrosomia)(aOR 3.01,aOR 2.12;aOR and 1.96;,respectively);in reference to non-diabetic women.Women with GD-NI were at a significantly increased risk for macrosomia(aOR1.53),decreased risk for their baby to die before their first birthday(aOR 0.41)and no difference in risk for preterm birth in reference to non-diabetic women.

The authors have no financial or non-financial competing interests or conflicts of interests associated with this manuscript.

The data used in this manuscript were acquired from the Missouri Department of Health and Senior Services(MODHSS)and are not available for public access due to MODHSS guidelines.The contents of this document including data analysis,interpretation or conclusions are solely the responsibility of the authors and do not represent the official views of DHSS.

The authors have read the STROBE Statement,and the manuscript was prepared and revised according to the STROBE Statement.

The sea lay before him, clear and blue, and the mussel shellsand pebbles, the playthings of his childhood, crunched over hisfeet. While he thus walked on his nose suddenly began to bleed; it was a trifling occurrence, but trifles sometimes are of greatimportance. A few large drops of blood fell upon one of his sleeves.

This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers.It is distributed in accordance with the Creative Commons Attribution NonCommercial(CC BYNC 4.0)license,which permits others to distribute,remix,adapt,build upon this work non-commercially,and license their derivative works on different terms,provided the original work is properly cited and the use is noncommercial.See:https://creativecommons.org/Licenses/by-nc/4.0/

United States

At last the young man gave in, and threw the fish back into the sea; and both brothers went supperless to bed, and wondered what fortune the next day would bring

Pamela K Xaverius 0000-0001-8824-2933;Steven W Howard 0000-0003-0056-6680;Deborah Kiel 0000-0003-0081-1252;Jerry E Thurman 0000-0003-3752-816X;Ethan Wankum 0000-0001-8470-1688;Catherine Carter 0000-0002-3839-4346;Clairy Fang 0000-0002-3570-2112;Romi Carriere 0000-0002-7157-4576.

Ma YJ

A

Ma YJ