Selective Screening for Gestational Diabetes Mellitus in Adolescent Pregnancies

MARY L. KHINE, MD, AMY WINKLESTEIN, AND JOSHUA A. COPEL, MD
Objective: It is unclear whether it is cost-effective to universally screen adolescent gravidas for gestational diabetes mellitus (GDM). Our objective was to identify the prevalence of gestational diabetes in our adolescent population and to review risk factors that would most easily identify a subset of adolescent patients at greatest risk for the development of gestational diabetes. Methods: Six hundred thirty-two adolescents were identified from 11,486 deliveries in our institution through International Classification of Diseases (9th edition) codes. Eleven of those adolescents had GDM. Their charts and those of a representative group of nondiabetic adolescent gravidas were analyzed for GDM risk factors, including race/ethnicity, body mass index (BMI), family history of diabetes, other medical disorders, and previous history of GDM, macrosomia, stillbirths, or anomalous fetus. Statistical analyses used ␹2 and Student t tests as appropriate. Results: The prevalence of GDM among adolescent gravidas was 1.7%. No difference was identified between the adolescent pregnancies with GDM and the comparison group for race/ethnicity, family history, or presence of medical disorders. Risk factors requiring a previous obstetric history could not be evaluated adequately because of the high prevalence of nulliparas. There was a higher prevalence of BMI over 27 in adolescents with GDM (P < .001). Conclusion: Body mass index is an important risk factor for the development of gestational diabetes in adolescent gravidas. We recommend that selective screening for GDM of adolescent gravidas be performed on the basis of BMI. (Obstet Gynecol 1999;93:738 – 42. © 1999 by The American College of Obstetricians and Gynecologists.)

Screening for gestational diabetes mellitus (GDM) has become standard practice in obstetrics. It was initially instituted to detect women who developed glucose intolerance during pregnancy,1 presumably because of the influences of human placental lactogen, cortisol, somatomedins, and other diabetogenic hormones altered by pregnancy.2 O’Sullivan and Mahan1 proposed

From the Section of Maternal Fetal Medicine, Yale University School of Medicine, New Haven, Connecticut.

criteria for the diagnosis of GDM in 1964 in an effort to identify women at risk for subsequently developing overt diabetes. These criteria have been adapted to identify women at risk for adverse perinatal outcome.3 Although GDM screening has become standard practice, it is not known how adverse perinatal outcome associated with GDM has been affected by its identification or treatment. Consequently, the topic of how to screen and whom to screen has been an area of unresolved debate. The ACOG recommended selective screening for GDM based on the presence of maternal risk factors for patients less than 30 years of age.4 The Third International Workshop and Conference on Gestational Diabetes and the American Diabetes Association recommended universal screening of all pregnant women.5 In support of the latter position, several investigators reported that up to 50% of women with GDM would have been missed if they were screened only on the basis of clinical or historic risk factors.6 – 8 Despite ACOG recommendations, 97% of responding residency directors and Maternal Fetal Medicine fellowship directors advocated universal screening.9 Similarly, more than 75% of recent residency graduates and ACOG Junior Fellows practiced universal screening themselves.10 Whether universal screening leads to fewer clinically significant perinatal complications is unknown. O’Sullivan and Mahan1 that perinatal mortality was not significantly higher in women who were slim and under the age of 25 years. Because the incidence of GDM increases with maternal age, those at the youngest reproductive ages are at lower risk for GDM. This suggests that the ideal patients for selective screening are those in low-risk populations, such as in adolescents.11 The American Diabetes Association recently published new recommendations for GDM screening.12 They now state that selective screening along clinical guidelines is acceptable for patients younger than 25

738 0029-7844/99/$20.00
PII S0029-7844(98)00550-X

Obstetrics & Gynecology

Results The general obstetric population at Yale New Haven Hospital is 70.years old. in search of the diagnosis of gestational diabetes or abnormal glucose tolerance and vaginal or cesarean delivery. The definition of teenage pregnancy is varied. 2 hours. 17. and obstetric history were recorded. MAY 1999 Khine et al Selective Screening for GDM 739 . Table 1. These resources might better benefit programs that are designed to deter the problem of teenage pregnancy. 1997 were scanned. There are few data on the most efficient way to screen selectively.486 Materials and Methods In our institution. Gestational diabetes mellitus might not be a clinically significant contributor to maternal or perinatal morbidity or mortality in adolescents and therefore might not warrant the resources consumed by universal screening. Statistical analysis was done using the ␹2 test and Fisher exact test where appropriate. Significant intrapartum events were noted. Fingerstick glucose values were evaluated at least once weekly. 93. as the prevalence of GDM in our adolescent population was unknown before the study. Although race/ethnicity. re- ported height at the initial prenatal visit (or reported prepregnancy values if first prenatal visit values were unavailable). Birth weight.3% white. length of stay in nursery. 32 control subjects for each case of GDM were analyzed. were also collected and stratified by age group. Insulin was given when fasting glucose values were persistently greater than 105 mg/dL or 2-hour postprandial values were persistently greater than 120 mg/dL. greater than 140 mg/dL. The control group consisted of 352 adolescent gravidas who were part of a previously existing database of adolescent patients who presented for obstetric care between January 1995 and January 1996. NO. such as maternal age. 5. most staff physicians practice universal screening of obstetric patients between 24 and 28 weeks’ gestation with a 50-g oral glucose load. Cutoff values were established by Carpenter and Coustan13: fasting. and adverse perinatal outcome of the infants were collected. Appropriate sample size was calculated after data were obtained. The remainder of the nondiabetic adolescent gravidas admitted between October 1995 and July 1997 were not reviewed because the number of existing cases within the database was sufficient to grant an ␣ of .5% Hispanic. It was believed to be appropriate for the present study because it included every adolescent gravida registered to the hospital in 1995. unequal sample sizes were required. 1995 and June 30. Racial and Age-Related Demographics Age (y) Ͻ19 Ն19 Total White 198 7849 8047 Black 301 1739 2040 Hispanic 129 853 982 Asian 3 184 187 Other 1 229 230 Total 632 10. Racial distribution based on maternal age is given in Table 1.9% other. As the prevalence of cases was low. There was no statistical difference of the incidence of GDM within racial/ethnic groups when stratified by payor status. As socioeconomic and perinatal complications exist more frequently in women younger than the age of 19 years. greater than 180 mg/dL. Demographic data for diabetic and nondiabetic patients. The Yale Human Investigation Committee approved the protocol.14 we included gravidas less than 19 years old at the time of prenatal care registration.8% black. All International Classification of Disease (9th edition) codes for hospital discharges between October 1. and family and obstetric history have been clinically useful as risk factors. family history. Although the statistical power was not significantly improved by increasing the number of control patients. Of 11. The only subjects that we excluded from analysis were those with pregestational or gestational diabetes. We performed this study to identify appropriate risk factors for gestational diabetes in adolescent gravidas. 8. and 1. according to Carpenter and Coustan. These guidelines are expansive. greater than 155 mg/dL. Subjects were identified by searching the hospital database for women admitted as of October 1. 1995. 1 hour. evidence of hypoglycemia.854 11. The prevalence of gestational diabetes increases with VOL.05 and a ␤ of . body habitus.8%) were complicated by gestational diabetes. 556 (4. The patient’s weight. Registered dietitians counseled women with GDM and placed them on a 30 –35 kcal/kg per day diet. Women were considered to have gestational diabetes if there were at least two abnormal values.486 deliveries at our institution over the study period. This database had been established as a reference group for another investigation. greater than 95 mg/dL.13 For patients with a positive screen.6% Asian. medical history. The test is performed without regard to time or content of previous meal.80 to detect a difference of 20% in mean body mass index (BMI) in light of the unequal sample sizes (nine study patients recommended). and 3 hours. PART 1. and payor status. a 100-g glucose load was administered after an overnight fast. 1. A positive screen was defined as a 1-hour postload value of at least 130 mg/dL of venous plasma. race. they might not be uniformly reliable in every patient population.

2% of gestational diabetics. PIH Cesarean for breech NSVD. only 16% had BMI greater than 27 compared with 82% in the gestational diabetic group (P Ͻ . shoulder dystocia NSVD.7 23 Birth weight ICU (g) stay 4190 2870 3765 2940 3720 4350 3340 4410 4050 3550 3660 Yes‡ No No No No No No No No No Yes‡ Delivery NSVD. Of the 11 adolescents with gestational diabetes.7% of adolescents. chlorioamnionitis NSVD.09% of our obstetrics population. PIH NSVD Cesarean for labor arrest.92 compared with 23.7%. and neonate 11 was admitted for 6 hours to treat symptomatic hypoglycemia. Prevalence of Gestational Diabetes by Age and Race Age (y) Ͻ19 19 –35 Ͼ35 White 1% (2/198) 4.3% (7/301) 5.2% (269/6457) 7.486 Table 4. Prevalence of Gestational Diabetes Mellitus by Age Group Age (y) Ͻ19 19 –24 25–35 Ͼ35 Total GDM 11 (1. except for patients 7 and 11. and 0. one for secondary arrest of labor at 6 cm (birth weight 4355 g) and the other for breech presentation (birth weight 3340 g).52 Ϯ 4. Asian women were more likely to have gestational diabetes overall.4 36 33 30.9%) 121 (7.2. compared with the racial representation of 70. Four women (36%) reported a history of diabetes in a first. Mild shoulder dystocia NSVD NSVD. The prevalence of GDM in subjects less than 19 years old was 1. According to the 1985 National Institutes of Health Consensus Panel on ObeTable 3. which represents nearly one standard deviation above the mean. Among nondiabetic adolescents. Of the remaining pregnancies. PIH NSVD.15 In the study population.4% (9/205) 8. obstetric history could not be evaluated.Table 2. The prevalence of gestational diabetes by age group is summarized in Table 2.001).1% (11/154) Hispanic ICU ϭ intensive care unit.3% (2/24) Total 1.4% in those over 35 years old (P Ͻ . The mean Ϯ standard deviation BMI of the gestational diabetics was 29.5% (2/129) 5.3% (43/806) 8.4%) 349 (4.5%). and a clear relationship between age and likelihood for GDM exists.2 29. n/a ϭ not available.7%) 75 (3. Excluding body habitus and race. compared with 21% in the comparison group (P ϭ . Prenatal charts for patients 7 and 11 were not available. Racial and age-related differences in patients with gestational diabetes are summarized in Table 3. with the most pronounced increase in the odds ratio among Asians. 4. seven (63%) were black.3% (15/162) 27. Asian 0% (0/3) 9. and 7.3% (6/22) Other 0% (0/1) 4. but none required long-term steroid medication. NSVD ϭ normal spontaneous vaginal delivery.8%) Non-GDM 621 2142 6649 1518 10.or second-degree relative. which were calculated on reported prepregnancy height and weight.001).930 Total 632 2217 6998 1639 11. Because 92% (10) of the adolescents in the GDM group and 88% of the adolescents in the comparison group (P ϭ nonsignificant) were nulliparous. nationally. 4th degree tear NSVD GDM ϭ gestational diabetes mellitus. nearly 80% of our adolescent population lacked historic risk factors for gestational diabetes.059). the 85th percentile was 27. Four reported other medical disorders (all asthma). advancing maternal age. ‡ Neonate 1 was admitted for a septic workup. In all adolescents.03 Ϯ 4. 632 were in mothers younger than 19 years old (5.2 based on their weight and reported height at initial prenatal examination or on reported prepregnancy height and weight (Table 4). * Gestational age at the time of prenatal care registration.7% (11/632) 4.68 in our comparison group. Two pregnancies were delivered by cesarean. One was delivered at 35 weeks due to preeclampsia. representing 1. 27. sity.0% (98/1392) Black 2.6 32 27.3% (P ϭ nonsignificant) in the general obstetric population. Perinatal Outcomes of Adolescents With Gestational Diabetes Estimated Body gestational mass index† age* Patient 1 2 3 4 5 6 7 8 9 10 11 12 16 19 12 8 12 n/a 9 20 5 n/a 28 36 21 27. Eleven of these patients were identified as having gestational diabetes.6% (88/1585) 7. PIH ϭ preeclampsia.4% (121/1639) 1. diabetes was controlled by diet alone. Salient features of these 11 pregnancies are given in Table 4. Ten of the eleven infants of diabetic adolescents were delivered at term. Of all deliveries during the study period. † Body mass index was calculated by kg/m2 based on weight and height at initial prenatal visit.4%) 556 (4. obesity may be defined as BMI greater than the 85th percentile.6% in those 19 –35 years old. Race becomes a more significant risk factor for the development of gestational diabetes as a patient ages. Nine of the adolescent subjects with GDM had a BMI greater than 27.3 for younger women.5% (4/47) 740 Khine et al Selective Screening for GDM Obstetrics & Gynecology .6% (424/9215) 7.

7% compared with the general population prevalence of 4. 82% of patients in whom GDM was eventually diagnosed by conventional screening would be identified. Although they found a statistically significant increase in the incidence of GDM with advancing age of the patient. The overall prevalence of gestational diabetes in our adolescent population was 1.3. Because most of the adolescents were nulliparous. house staff. Body mass index higher than 27 is useful in identifying adolescents at greater risk for having abnormal results of glucose challenge and tolerance testing. Because of the retrospective nature of the study. fetal anomalies. none of our small cohort of diabetic adolescents had diabetes postpartum. Although this is of concern. the negative predictive value is 99. increasing age. 93. and 2-hour values and maternal age. who remained an extra 4 days for completion of a septic workup (patient 1). Other risk factors frequently cited. including attending physicians. The mean birth weight of approximately 3710 g for the infants of diabetic mothers was not significantly different from the mean birth weight of 3090 g for the infants born to the comparison group (P Ͼ . they only included women who were at least 25 years old.059).6 – 8. parity.8%.19 but few uniformly examined both age and BMI.001). who may have less insulin reserve.3%). We recommend screening for gestational diabetes in adolescent gravidas with BMI greater than 27. previous macrosomic infant (43%). A prospective study might control for some of those variables. Discussion Of the classic risk factors. several nurses. Maternal BMI greater than 27 in nine (82%) of 11 adolescents with GDM contrasts with the overall prevalence of 16% in the comparison population (P Ͻ . In the present study. Family history of a firstor second-degree relative with diabetes might be a useful adjunct in selecting a subset for screening because the prevalence of a positive family history in our adolescent population approached statistical significance (P ϭ . prior obstetric history was not a useful factor. The prevalence of gestational diabetes in our general obstetric population was within the range reported in the literature (2.18.18 In 1989. particularly in younger patients.05). 5.17 Our data strongly support this association. nurse practitioners. Weeks et al7 reported that the most common risk factors in patients with gestational diabetes were family history of diabetes (45%). Only one infant had hypoglycemia requiring intravenous glucose (patient 11).0 – 6.4%. and glycosuria. 1-hour. Demographic differences in the two populations (our population consisted of a larger percentage of Asian and Hispanic patients) might have contributed to that difference.2 The reason for this may be due to our exclusion of older gestational diabetics. age and BMI have repeatedly been associated with the development of diabetes and glucose intolerance in pregnant women. Universal screening. All infants were discharged to the care of their mothers on the day of maternal discharge. none of them required insulin. such as presence of other medical disorders. Naylor et al21 recently proposed a clinical risk factor scoring system as a means of GDM screening. Some studies even excluded obese subjects from the analysis.7 Sacks et al20 reviewed plasma glucose values of 3505 pregnant women given a 75-g. with an overall GDM rate of 2%. whereas other studies have predicted that 15–17% of gestational diabetics would require insulin to maintain optimal glycemic control. and higher BMI were the most statistically significant risk factors for GDM. PART 1. as our small study group might not appropriately represent a larger population. could be supported by possible detection of undiagnosed pregestational diabetes. Other studies examined the importance of risk factors in detecting gestational diabetes. Race.16. 2-hour glucose tolerance test (World Health Organization recommendations) and found positive correlation between fasting. except for one. That infant was released home with the mother and did not require a prolonged stay in the nursery. The birth weights of these fetuses were 4190 g and 4410 g. VOL. although it is most likely because Coustan et al used the National Diabetes Data Group’s criteria to identify gestational diabetics. The absence of some risk factors in our study could be a function of recall bias (younger subjects frequently are unaware of complete family histories) and the prevalence of nulliparous women (88%) in this adolescent population. Four infants in the GDM group weighed more than 4000 g. We interpret the poor association of race/ethnicity and the development of gestational diabetes in our data with caution. and prepregnancy BMI. NO. we relied on data gathered by a large group of people with different levels of history-gathering expertise. When a cutoff BMI of greater than 27 was used. MAY 1999 Khine et al Selective Screening for GDM 741 .two neonates had mild shoulder dystocia without neurologic impairment reported on newborn examination. There were no cases of clinically apparent electrolyte disorders or hyperbilirubinemia. were not noted at an increased frequency in adolescents with GDM compared with controls. Additionally.7% in patients less than 20 years old. and students. age and race seem to be significant co-risk factors for older age groups but less important in the younger age group. Coustan et al18 reported a preva- lence of 0. and prior gestational diabetes (30%).

10. Yao JFF. were not helpful in our young population. a case-control design was most appropriate. et al. MD Department of Obstetrics and Gynecology NYU Downtown Hospital 170 William Street.269:609 –15. The overall prevalence of gestational diabetes is low in the adolescent population.306:37– 8. Received in revised form October 26. Diabetes in pregnancy. or presence of other medical disorders. Mahan CM. Am J Obstet Gynecol 1073. making selective screening attractive. Knops RH. 8th Floor New York. Mary L. Magee MS.23:125–36. Wolde-Tsadik G. 5. Rodriguez H. Charles D. However. Third International Workshop Conference on Gestational Diabetes Mellitus. Hankins GDV. 5% had 1-hour postglucose load values between 130 and 139. JAMA 1993. Benedetti TJ. Screening and testing for gestational diabetes mellitus. Coustan DR.21:s1–9.144:768 –73. Stamford. Landon MB. 3. Khine. 13. O’Sullivan JB. 742 Khine et al Selective Screening for GDM Obstetrics & Gynecology . 1997. O’Sullivan JB. Carpenter MW. 20. 2. conclusions concerning large demographic variables such as race are made with caution as just one or two additional subjects in the study group could influence the statistical analysis. 15. Maternal fetal medicine: Principles and practice. 18. Published by Elsevier Science Inc. Cunningham FG. In: Creasy RK. eds. ACOG technical bulletin no. Maternal age and screening for gestational diabetes: A population based study. Copyright © 1999 by The American College of Obstetricians and Gynecologists.116:895–900.13:642–5. Henry H. Toward universal criteria for gestational diabetes: The 75 gram glucose tolerance test in pregnancy. 20th ed. Greenspoon JS. Report of the expert committee on the diagnosis and classification of diabetes mellitus. Coustan DR.8:125–7. Leveno KJ. The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. 40(Suppl):197–201.171:1003–7. Krohn H. Am J Obstet Gynecol 1995. gestational diabetes mellitus. Teenage pregnancy: solutions are evolving. Jarrett RJ. Connecticut: Appleton & Lange. Screening criteria for high risk gestational diabetic patients. We did not find race/ethnicity helpful in identifying patients more likely to have diabetes during pregnancy. Gilstrap LC III. Am J Obstet Gynecol 1973. Dandrow RV. Screening adolescent gravidas for gestational diabetes. Influence of diagnostic criteria on the incidence of gestational diabetes and perinatal morbidity. Moore TR. Washington DC: American College of Obstetricians and Gynecologists. Adoles Pediatr Gynecol 1995.337:1591–6. Am J Obstet Gynecol 1982. American College of Obstetricians and Gynecologists. 12. 3rd ed. Management of diabetes in pregnancy. 9. 4. Carpenter MW. 14. Diabetes 1991. Kreutner AK. Sacks DA. Gestational diabetes: A non-entity? BMJ 1993. O’Sullivan JB. Major CA. suggesting that a BMI of 27 or more might be the most useful criterion for identifying patients most likely to benefit from diabetes screening. Carr SR. Summary of recommendations of the third international workshop conference on gestational diabetes. Owen J. Williams obstetrics. Traditional risk factors. Obstet Gynecol 1989. 1998. 17. Widness JA. 1994:934 –77.172:615–20. Chen E. Neiger R. Coustan DR. Because of the low prevalence of gestational diabetes in this age group. Family history of diabetes was not found statistically more frequently in the study population compared to the control population. Dandrow RV. this design also represents the largest weakness to this study. Phelan ST. List M.73:557– 60. Criteria for the oral glucose tolerance test in pregnancy. Charles D.172:607–14. American Diabetes Association. N Engl J Med 1980. Diabetes Care 1998. although there was a trend toward statistical significance. 19. Gestational diabetes and perinatal mortality rate. Diabetes Care 1998. Resnik R. Management of diabetes mellitus and pregnancy: A survey of obstetricians and maternal fetal medicine specialists. 7. Hollingsworth DR. Diabetes 1964. 1998. Gabbe SG. Clinical practice recommendations 1998 —Position statement. 21. Nelson C. as the conclusions were based on record review of a small number of cases. Farine D (for the Toronto Tri-hospital Gestational Diabetes Project Investigators). 6. Am J Obstet Gynecol 1995.211:s5–19. Naylor CD. 1994. Gant NF. Sermer M. Criteria for screening tests for gestational diabetes. We increased the power to detect a difference between the control and study groups by using a large number of controls for each case of adolescent GDM. We believe that this result is not significant because the control population was evaluated using the same criteria and the evaluation of risk factors for GDM was independent of the cutoffs.75:635– 40. 16.117: 901–4. Gabbe SG. Thompson S. 11. Obstet Gynecol 1990.The incidence of GDM in any given population could be 15% higher using Carpenter and Coustan’s screening and diagnostic cutoffs compared with those proposed by ACOG and the National Diabetes Data Group. Morgan MA. which indicates that the incidence of GDM in our population might be overestimated. obstetric history. MacDonald PC. Accepted October 29. Abu-Fadil S. Sachs L. Address reprint requests to: References 1. Mahan CM. Selective screening for gestational diabetes mellitus. We could not determine whether perinatal outcome was improved with treatment of GDM in adolescent pregnancies in the present study because the number of diabetic adolescents was too small to draw significant outcome conclusions.303:516 – 8. such as personal history of diabetes. Rotondo L. 1998. Sounding board. Obstet Gynecol Clin North Am 1996. Gestational diabetes survey. Walden CE. Nine of eleven adolescent gravidas with GDM reported early pregnancy BMI greater than 27. Philadelphia: WB Saunders. Weeks JW. 8. Nonetheless.13 In the 1-hour glucose challenge tests in our study population. Landon MB. 200. N Engl J Med 1997. NY 10038 Received June 29. The importance of examining all variables in this study is the acknowledgment of trends. Gestational diabetes: Does the presence of risk factors influence perinatal outcome? Am J Obstet Gynecol 1994. deVeciana M. Mahan CM.

Sign up to vote on this title
UsefulNot useful