Characteristics of Gestational Diabetic Mothers andTheir Babies in An Indian Diabetes Clinic

Shailaja D Kale, SR Kulkarni, HG Lubree, K Meenakumari, VU Deshpande, SS Rege,J Deshpande, KJ Coyaji, CS Yajnik

INTRODUCTION

Gestational Diabetes Mellitus (GDM) is defined asglucose intolerance that has its onset or firstrecognition during pregnancy.1 GDM is common inpopulations who have high rates of type 2 diabetes.Common risk factors for GDM include familialpredisposition, higher age, overweight, and previoushistory of large babies.2 Recently, small birth weight andshort height (of the pregnant woman) have also beenshown to increase risk of GDM.3 India has the highestnumber of patients with type 2 diabetes in the world4 but only sparse information is available on theprevalence, predictors and the risks of GDM in India.

Diabetes Unit, King Edward Memorial Hospital and Research Center, Rasta Peth, Pune-11, India.

Received : 30.11.2004; Revised : 28.2.2005;Re-revised : 2.5.2005; Accepted : 4.6.2005

GDM increases the risk of bad outcome in thepregnancy for both the mother and the baby, and alsoincreases the risk of ‘permanent’ diabetes for themother.5-7 GDM mothers have an excess of metabolic riskfactors, which qualify them for the diagnosis of themetabolic syndrome and an increased cardiovascularrisk.8,9 We have reported on the high incidence of type 2diabetes and metabolic syndrome 4 years after deliveryin GDM mothers in our clinic.10 In this present study wecompare the characteristics of GDM mothers treated inour Diabetes Clinic, their pregnancy outcome and thecharacteristics of their babies with those of non-diabeticmothers and their babies treated in the same hospital.

MATERIALS AND METHODS

We studied retrospectively women with gestationaldiabetes mellitus (GDM) treated in Diabetes Unit, KingEdward Memorial (KEM) Hospital, Pune, India duringthe period Jan 1998 to December 2003 to compare thecharacteristics of GDM pregnancies with a non-diabeticgroup. We collected comparable information from non-diabetic mothers delivered in the KEM Hospital duringFeb-May 1998. The KEM Hospital provides obstetricservices to the local population and also providessecondary and tertiary services for women in Pune cityand surrounding villages. Approximately 1500 womendeliver every year in this hospital. There is no universalscreening for diabetes in pregnancy. Women thought tobe at high risk of diabetes by the treating obstetricianundergo a 75 g OGTT; these include: women with firstdegree relatives with diabetes, bad obstetric history,overweight, excessive weight gain during pregnancy,those with glycosuria or a high random plasma glucose,hypertensive, ultra-sonographic diagnosis ofmacrosomic baby, polyhydramnios or a congenitalanomaly.

GDM was diagnosed by WHO-1998 criteria;11 bothimpaired glucose tolerant and diabetic women wereconsidered as having gestational diabetes. They weretreated as per an agreed protocol. Dietary advice includedsugar restriction, increased fiber rich foods and frequentsmall feeds rather than two major meals. Exercise advicewas in consultation with the obstetrician, for mostwomen it was walking for at least 20 minutes daily. Ifplasma glucose concentrations were still elevated after1 week of dietary adjustment (fasting plasma glucose>95 mg/dl and 2h post meal > 140 mg/dl), insulintreatment and home monitoring of blood glucose wasadvised. Obstetric management followed the Hospitalprotocol for high-risk pregnancies. This included serialsonography to date the pregnancy, diagnose fetalanomalies, and monitor fetal growth. Non-stress testing(NST) was performed after 34 wks of gestation. Treatingobstetrician decided appropriate time and mode ofdelivery. Informed consent was obtained from thepatients for anthropometric and blood measurementsand the Hospital Ethical Committee approved the study.

Measurements

Maternal information during pregnancy was obtainedfrom outpatient records. This included medical andobstetric history, the time and the results of the glucosetolerance test, and the time and the mode of delivery.Gestational age at delivery was calculated from the lastmenstrual period, supported by sonographic findings.Maternal venous blood sample was collected soon afterdelivery.

Anthropometry

Standardized anthropometric measurements weremade in duplicate on the left side by two trained observers within 24 hrs of delivery. Mean of two measurementswas used in analysis. Maternal measurements includedweight using Soehnle electronic scales, height usingHarpenden Stadiometer; biceps, triceps, subscapular,and suprailiac skinfolds using Harpenden skinfoldcalipers; and head, midarm, waist and hipcircumferences using a non-stretchable fiberglassmeasuring tape. All the anthropometric instrumentswere supplied by CMS Instruments, London, UK. Bloodpressure was recorded in sitting position using amercury sphygmomanometer.

Neonatal measurements included: weight usingATCO scale (Wadala, Mumbai, India); crown-heel lengthusing Pedobaby; subscapular and triceps skinfolds onthe left side of the body using Harpenden skinfoldcalipers and head, mid-upper arm, abdominal, and chestcircumferences using a non-stretchable fiberglassmeasuring tape. Placenta was weighed soon after birthafter cutting the cord flush and trimming off themembranes.

Maternal blood measurements

Maternal venous blood was collected soon after thedelivery. Hematological measurements were made on aCoulter analyzer AcT diff TM (Beckman, Miami, FL, US).Plasma measurements included: glucose, albumin, totaland high-density lipoprotein (HDL) cholesterol andtriglycerides. These were performed using standardclinical laboratory methods on a Spectrum BiochemistryAnalyzer (Abbott Laboratories, Irwing, TX, US).

Controls

We collected similar information at delivery inmothers without diabetes who delivered in the KEMHospital during the period Feb 1998 to May 1998. Thosewho delivered full term with uncomplicated pregnanciesand had fasting plasma glucose concentrations <90 mg/dl and 2h post meal < 120 mg/dl during pregnancywere selected as controls.

Statistical Methods

Data is shown as mean (SD) for normally distributedvariables and as median (IQR) for those not normallydistributed. The latter were log transformed to ensurenormality during statistical analysis. Groupcomparisons were made using chi-square test. Thesignificance of the difference between means of the twogroups was analysed by ANOVA with adjustments forconfounding variables as appropriate. Multiple linearregression was used to determine the effects of maternalcharacteristics on the size of the newborn. Analysis wascarried out using STATA (version 7.0) (Stata Corporation,TX, USA).

RESULTS

Pregnancy characteristics

GDM

Three hundred and thirty three pregnant diabetic women were referred to the Diabetes Unit formanagement during the period Jan 1998 to Dec 2003.Sixty-five women had pre-gestational diabetes (20 type1, 42 type 2 and 3 fibro-calculous pancreatic diabetes)and 268 had gestational diabetes. Three womendelivered twins. Further analysis therefore refers to 265singleton gestational diabetic pregnancies.Anthropometric measurements were available on 163mothers (Fig. 1) and 247 neonates; post-delivery bloodresults were available on 128 women.

Mean gestation at diagnosis of GDM was 32 wks(range 26-35 wks), 150 women had IGT and 115 haddiabetes by the WHO 1998 criteria. Eighty six percentmothers were primiparous. One hundred and twentyeight (48%) mothers had a first-degree relative withdiabetes. Mean gestation at delivery was 38 wks, 61women (23%) delivered pre-term (<37 wks). One hundredand fifty-eight women (60%) were delivered by caesareansection. There were 4 intrauterine deaths (IUD) and 56babies were diagnosed with a neonatal complication (2respiratory distress, 20 hypoglycemia, 12 jaundice, 2sepsis, 5 congenital anomalies, 17 other).

One hundred and sixty-two women (61%) receivedinsulin treatment and 103 (39%) women were treatedwith diet and exercise. Insulin treatment was started ata mean gestation of 28 weeks. Median insulin dose was20 units/d (8-105 units/d), 14 women received >1unit/kg/d. Oral hypoglycemic agents were not used butmajority of women were prescribed guar-gum (Carbotard,Jagat Pharma, Pune).

Controls

Of 463 non-diabetic mothers delivered in the KEMHospital during Feb-May 1998, 315 were residents ofPune. Two hundred and fifteen delivered full term (= 37week of gestation). Anthropometric measurements wereavailable on 187 mothers and 192 neonates, and post-delivery blood results on 121 women.

Of 215 non-diabetic mothers 68.0% were primiparous.Mean gestation at delivery was 39 wks. Sixty-six (31.0%) were delivered by caesarean section. There were no IUDs,and none of the babies was clinically diagnosed withany complications. or congenital anomalies.

Anthropometry and blood pressure:

(Table 1) Because of significant differences between GDM andnon-diabetic mothers all anthropometric and bloodmeasurement comparisons are adjusted for age, parityand gestation at delivery. GDM mothers were heavier,more obese (higher BMI) and more centrally obese (higherWHR) compared with non-diabetic mothers. All skinfold thicknesses were significantly higher in GDMmothers (adiposity) including in mothers who had BMI<25 kg/m2. GDM mothers had significantly higher bloodpressure.

Blood measurements:

(Table 2)GDM mothers had higher concentration of bloodhemoglobin, higher hematocrit and higher platelet countbut not red cell and total leucocyte count compared tonon-diabetic mothers. GDM mothers also had higherconcentrations of plasma triglycerides and lowerconcentrations of plasma HDL cholesterol compared tothose in the non-diabetic mothers. This was also true inmothers who had a BMI <25 kg/m2 (Data not shown).

Metabolic syndrome variables:

(Table 3)GDM mothers had higher prevalence of overweightand obesity (BMI>25 kg/m2), central obesity(WHR>0.85), hypertriglyceridemia (plasmatriglyceride>150 mg/dl) and hypertension (bloodpressure>140/90 mm Hg) compared to the non-diabeticmothers.

Comparison of mothers with IGT and DM

Diabetic mothers were older than those with IGT (31vs. 29y, p<0.01). Both groups had similar body sizemeasurements and other biochemical parameters.Mothers with diabetes had higher systolic bloodpressure (p<0.01). Sixty four percent of diabetic mothersand 54% of IGT mothers underwent caesarean section(p=ns).

Comparison of insulin treated and diet treated GDMmothers:(Table 4)

Insulin treated mothers were older (31 vs. 28y,p<0.001), and more adipose (sum of four skinfolds 105.2 vs. 87.4mm, p<0.001). They had higher plasma glucoseconcentrations during diagnostic GTT (p<0.001), werediagnosed at an earlier gestation (p<0.05) and alsodelivered one week earlier (37.7 vs. 38.4wks, p<0.05).There was no significant difference in BMI, WHR, headcircumference and blood pressure in the two groups.

Offspring characteristics: (Table 5)

Babies of GDM mothers were delivered one weekearlier than the babies of non-diabetic mothers. Allcomparisons were corrected for gestation and gender. Inaddition GDM mothers were larger in size and haddifferent parity therefore the neonatal comparison werefurther adjusted for maternal BMI and parity. Babies ofGDM mothers were heavier (2936.0 vs. 2809.0g, p<0.001)compared to those born to non-diabetic mothers whenadjusted for gender, gestation at delivery and maternalage and parity. This difference did not remain significantwhen adjusted additionally for maternal BMI or height.Thirty percent of babies of GDM mothers weighed morethan 90th centile of birth weight of babies of non-diabetic mothers (3269.0g). Five percent of babies of GDM mothersand 0.6% of babies of non-diabetic mothers hadmacrosomia (weight > 4000.0g).

Babies of GDM mothers were longer, had largermidarm and abdominal circumferences and subscapularand triceps skinfold thicknesses. There was nosignificant difference in the ponderal index and headand chest circumferences. Placenta of the babies of the GDM mothers was also heavier compared to the placentaof the babies of non-diabetic mothers. There was nosignificant difference in any of the anthropometricmeasurements of babies born to IGT or diabetic mothersor between those treated with insulin or diet.

In GDM pregnancies greater maternal height andlarger waist circumference predicted higher birth weightof the baby (adjusting for gestation at delivery andgender). However in non-diabetic pregnancies highermaternal weight, head circumference, waistcircumference and mid-arm were associated with higherweight of the baby.

In GDM pregnancies, maternal fasting and 2h plasmaglucose concentrations at the time of OGTT did notpredict birth weight of the babies, maternal biochemicalparameters at delivery were also not related. In non-diabetic mothers, higher hematocrit, plasma cholesterol,triglycerides and albumin concentrations wereassociated with higher weight of the babies.

Mothers of the babies who had any complication atbirth were diagnosed diabetic earlier in the pregnancy(27.5 vs. 30.5wks, p<0.04) and also delivered one weekearlier (36.9 vs. 38.3wks, p<0.001). There were no otherpredictors of neonatal complications.

DISCUSSION

The results confirm the premise that excess oftraditional risk factors for type 2 diabetes (family historyof diabetes and obesity) and cardiovascular diseaseoccur in Indian GDM mothers as compared to non-diabetic mothers. This suggests that pregnancy is awindow to future cardiovascular risk.9 Babies of GDMmothers were larger and more adipose compared tobabies of non-diabetic mothers.

Using WHO criteria only a third of these mothers haddiabetes while two thirds had IGT. Despite this, one week after nutritional treatment, sixty percent mothersexceeded the target glycemic levels for good control andwere put on insulin treatment. Incidence of maternalmorbidity was substantial including pre-term deliveryand caesarean section. The rate of caesarean section wassimilar in mothers with IGT or diabetes, but insulintreated women had higher rates. Insulin treated motherswere older, more adipose and not unexpectedly had moresevere diabetes which was diagnosed at an earliergestation.

Even though the GDM mothers were older than thenon-diabetic mothers, the mean age of GDM mothersunder study was only 29 years, and the youngest GDMmother was 20y old. Thus, Indian GDM women areyounger compared to white Caucasian GDM women(mean age 33 years).6 This reflects the tendency of Indiansto acquire type 2 diabetes at a younger age whencompared to white Caucasians.12,13 The most strikingphysical characteristic of the GDM mothers was theirincreased obesity as compared to the non-diabeticmothers. Even those who were ‘normal’ by current BMIcriteria (<25 kg/m2, WHO 2004)14 were more adipose(higher body fat percent). Indians are reputedly moreadipose for a given BMI compared to other races andethnic groups.15-18 In this clinic-based study we were notable to confirm the previous observation of smaller heightas predictive of GDM.

GDM mothers in our study showed an excess ofcardiovascular risk factors mostly related to insulinresistance and the metabolic syndrome (obesity,adiposity, central obesity, higher plasma triglycerideconcentration and higher blood pressure). The elevatedblood hemoglobin concentration in the GDM mothers isnot easy to interpret. It could indicate a relatively smallervolume expansion in GDM mothers (an unfavorablefactor for fetal growth) or it could reflect better iron statusdue to better dietary intake or supplementation in a moreclosely supervised pregnancy. Higher platelet count inthe GDM mothers is suggestive of a prothrombotic state,and with higher blood pressure suggests endothelialdysfunction.19 Thus, these findings during pregnancyin GDM Indian mothers anticipate the heightened riskof metabolic syndrome and cardiovascular risk, whichwe have reported in these mothers 4y after the delivery.10

Despite earlier delivery, mean birth weight of babiesof GDM mothers was higher than that in those of non-diabetic mothers. And the difference was predominantlyin the soft tissues (adiposity, abdominal circumferenceand midarm circumference); skeletal measurementsexcept height were similar. This observation fits wellwith the general concept of stimulated insulin dependentgrowth in babies of diabetic mothers.20 It is relevant tonote that only 5 percent of the babies of GDM motherswere ‘macrosomic’ by international definition (>4000g).However, a third of the babies of GDM mothers wereheavier than 90th centile for babies of non-diabetic mothers. Caution should be exercised when usinginternational criteria to describe morbidity in Indians.

Maternal glycemia at the diagnosis of GDM did notpredict baby’s birth weight, nor did maternal circulatinglipids and albumin concentrations at delivery. This maypartly be attributable to the phenotype modification bythe anti-diabetic treatment. On the other hand, highercirculating cholesterol and triglyceride concentrationsin non-diabetic mothers predicted larger birth weight inthe babies.

The information on the neonatal morbidity wasobtained retrospectively form hospital records; thereforewe have to interpret this with caution. Babies of GDMmothers had higher neonatal morbidity (hypoglycemia,respiratory distress, sepsis etc). The only predictors ofneonatal morbidity were related to earlier delivery andearly gestation at diagnosis of GDM, which indicatesthat some of these mothers may have been diabetic priorto pregnancy. Ours is a hospital-based study where only‘high risk’ mothers are screened for diabetes. Thus,admittedly there is a likely bias in our sample. On theother hand our data represents the current clinicalpractice in a large teaching hospital. A prospective studywith universal glucose tolerance testing prior topregnancy would provide a more precise idea of the riskfactors for GDM, as well as the risk to the pregnancyand the offspring.

In summary, the result shows that gestational diabetesmellitus occurs in relatively young urban Indian motherswho have a family history of type 2 diabetes and areadipose. Some of these women may have diabetes priorto pregnancy. These young women have metabolicabnormalities, which suggest increased cardiovascularrisk. Obstetricians should practice universal screeningfor diabetes in urban mothers to improve the outcomeboth for the baby and the mother. There is a need forlarge controlled studies to confirm the risks of GDM forIndian mothers and their babies.

REFERENCES

1.Metzger BE, Coustan DR. (Eds): Proceedings of the FourthInternational Workshop- Conference on Gestational DiabetesMellitus. Diabetes Care 1998;21(Suppl 2):B1-B167.

2.American Diabetes Association. Gestational diabetesmellitus. Diabetes Care 2003;26: s103-s105.

3.Egeland GM, Skjaerven R, Irgens LM. Birth characteristics ofwomen who develop gestational diabetes: population basedstudy. BMJ 2000;321:546-47.

4.Wild S, Roglic G, Green A, Sicree R, King H. Global prevalenceof diabetes estimates for the year 2000 and projections for2030. Diabetes Care 2004;27:1047-53.

5.Damm P, Kuhl C, Bertelsen A, Molsted-Pedersen L. Predictivefactors for the development of diabetes in women withprevious gestational diabetes mellitus. Am J Obstet andGynecol 1992;167:607-16.

6.Coustan DR, Carpenter MW, O’Sullivan PS, Carr Stephen R.Gestational diabetes: Predictors of subsequent disorderedglucose metabolism. Am J Obstet and Gynecol1993;168:139-45.

7.Kim C, Newton K, Knopp R. Gestational diabetes and theincidence of Type 2 diabetes. Diabetes Care 2002;25:1862-68.

8.Luis FP, Lucrecia H, Martin-Vaquero P, Garcia-Ingelmo T,Grande C, Janez M. Impaired fasting glucose and impairedglucose tolerance in women with prior gestational diabetesare associated with a different cardiovascular profile. DiabetesCare 2003;26:2318-22.

9.Sattar N, Greer IA. Pregnancy complications and maternalcardiovascular risk: opportunities for intervention andscreening. BMJ 2002;325:157.

10.Kale SD, Yajnik CS, Kulkarni SR, Meenakumari K, JoglekarAA, Khorsand N, Ladkat RS, Lubree H, Ramdas L. Highrisk of diabetes and metabolic syndrome in Indian womenwith gestational diabetes mellitus. Diabetic Med2004;21:1257-58.

11.Alberti KGMM, Zimmet PZ for the WHO Consultation.Definition, diagnosis and classification of diabetes mellitusand its complications, part I: diagnosis and classification ofdiabetes mellitus. Provisional report of a WHO Consultation.Diabet Med 1998;15:539-53.

12.UK Prospective Diabetes Study Group : UK ProspectiveDiabetes Study XII: differences between Asian, Afro-Caribbean and white Caucasian type 2 patients at diagnosisof diabetes. Diabetic Med 1994;11:670-77.

13.Mohan V, Alberti KGMM. Diabetes in the tropics. In AlbertiKGMM, Zimmet P, DeFronzo RA, Keen H (eds). Internationaltextbook of diabetes mellitus, second edition. John Wileyand Sons Ltd, 1997;172-87.

14.WHO Expert Consultation. Appropriate body mass indexfor Asian populations and its implications for policy andintervention strategies. Lancet 2004;363:157-63.

15.Banerji MA, Faridi N, Atluri R, Chaiken RL, Lebovitz HE.Body composition, visceral fat, leptin and insulin resistancein Asian Indian men. J Clin Endocrinol Metab1999;84:137–44.

16.Chandalia M, Abate N, Garg A, Stray-Gundersen J, GrundySM. Relationship between generalized and upper bodyobesity to insulin resistance in Asian Indian men. J ClinEndocrinol Metab 1999;84:2329–35.

17.Yajnik CS, Lubree HG, Rege SS, Naik SS, Deshpande JA,Deshpande SS, Joglekar CV, Yudkin JS. Adiposity andhyperinsulinemia in Indians are present at birth. J ClinEndocrinol Metab 2002;87:5575-80.

18.Yajnik CS, Yudkin JS. Y-Y paradox. Lancet 2004; 363:163.

19.Knock GA, McCarthy AL, Lowy C, Poston L. Association ofgestational diabetes with abnormal maternal vascularendothelial function. Br J Obstet Gynecol 1997;104:229-34.

20.Pedersen J, Molsted-Pedersen L. The hyperglycemia-hyperinsulinism theory and the weight of the newborn baby.In Rodrigues RR, Wallance-Owen J (eds). Diabetes. ExcerptaMedica,: Amsterdam. 1971;678-82.