- Email: [email protected]
Optimal Timing of Prenatal Ultrasound in Predicting Birth Weight in Diabetic Pregnancies
OBSTETRICS
Optimal Timing of Prenatal Ultrasound in Predicting Birth Weight in Diabetic Pregnancies Christopher M. Nash, MD, MSc;1 Christy Woolcott, PhD;1,2 Colleen O’Connell, PhD;1 B. Anthony Armson, MD, MSc1 1
Department of Obstetrics and Gynaecology, Dalhousie University, Halifax, NS
2
Department of Pediatrics, Dalhousie University, Halifax, NS
Abstract
Résumé
Objective: This study sought to determine the optimal timing of ultrasound in the third trimester to predict birth weight accurately in diabetic women with a singleton pregnancy.
tude visait a de terminer le meilleur moment pour effectuer Objectif : L’e chographie au troisie me trimestre chez les femmes une e tiques qui ont une grossesse monofœtale afin de pre dire avec diabe la naissance. exactitude le poids a
Methods: A retrospective cohort study of all diabetic women with a singleton pregnancy treated in Halifax, Nova Scotia, was performed. Estimated fetal weight was derived from ultrasound measures using the Hadlock2 equation. The Mongelli equation was used to predict birth weight. The association between gestational age at ultrasound and accuracy of predicted birth weight was assessed, with accuracy as a continuous variable representing the difference between predicted and actual birth weight and as a categorical variable (with four gestational age categories) representing whether predicted birth weight was within, over, or under 250 g of actual birth weight Results: The cohort of 943 women comprised 121 (12.8%) with type 1 diabetes, 111 (11.7%) with type 2 diabetes, and 711 (75.4%) with gestational diabetes. Ultrasound scans performed at term were the most accurate in predicting birth weight. At this gestational age, the mean difference between predicted and actual birth weight was 30 g (95% confidence interval 109 to 48). After adjusting for maternal body mass index, age, smoking, type of diabetes, and interval between ultrasound examination and delivery, accuracy improved as gestational age at ultrasound increased (P = 0.005). The odds of underpredicting or overpredicting birth weight were not significantly affected by the timing of the ultrasound examination. Conclusion: Because the predictive accuracy of ultrasound prediction of birth weight improves with gestational age, fetal growth assessment at term is recommended to aid with delivery planning in women with diabetes.
Key Words: Estimated fetal weight, diabetes, diagnostic accuracy, ultrasound, birth weight Corresponding author: Dr. Christopher Nash, Department of Obstetrics and Gynaecology, Dalhousie University, Halifax, NS. [email protected] Competing interests: See Acknowledgements. Each author has indicated that they meet the journal’s requirements for authorship.
alise une e tude de cohorte re trospective de toutes Méthodologie : On a re tiques qui connaissaient une grossesse monofœtale les femmes diabe taient traite es a Halifax, en Nouvelle-Ecosse. et qui e Le poids fœtal ae te calcule a partir des mesures e chographiques en utilisant estime pre dire le la formule de Hadlock 2. La formule de Mongelli a servi a la naissance. L’e tude a e value le lien entre l’a ^ ge gestationnel poids a chographie et l’exactitude du poids a la naissance au moment de l’e dit, ou la l’exactitude est une variable continue qui correspond a pre rence entre les poids a la naissance pre dit et re el, ainsi qu’une diffe gories d’a ^ge gestationnel) qui variable nominale (avec quatre cate termine si le poids a la naissance pre dit correspond au poids a la de el ou comporte une diffe rence de plus ou moins 250 g. naissance re Résultats : La cohorte de 943 femmes se composait de 121 femmes te de type 1, 111 femmes (11,7%) (12,8%) atteintes du diabe te de type 2 et 711 femmes (75,4%) atteintes du atteintes du diabe te gestationnel. Les e chographies effectue es a terme ont diabe cet a dit le poids a la naissance avec le plus de pre cision. A ^ge pre rence moyenne entre les poids a la naissance gestationnel, la diffe dit et re el est de 30 g (intervalle de confiance de 95% de 109 pre 48). Apre s un ajustement pour tenir compte de l’indice de masse a ^ge, le tabagisme, le type de diabe te et corporelle maternel, l’a chographique et l’accouchement, la l’intervalle entre l’examen e cision s’est ame liore e a mesure que l’a ^ge gestationnel au pre chographie augmentait (P = 0,005). Le moment de moment de l’e chographique n’a pas eu d’incidence notable sur les l’examen e s de sous-estimer ou sure valuer le poids a la naissance. probabilite diction du poids a la Conclusion : Puisque l’exactitude de la pre l’e chographie s’ame liore avec l’augmentation de l’a ^ge naissance a d’e valuer la croissance fœtale a gestationnel, il est recommande planifier l’accouchement des femmes atteintes terme pour aider a te. du diabe © 2019 The Society of Obstetricians and Gynaecologists of Canada/La Société des obstétriciens et gynécologues du Canada. Published by Elsevier Inc. All rights reserved.
Received on February 15, 2019
J Obstet Gynaecol Can 2019;000(000):1−6
Accepted on May 6, 2019
https://doi.org/10.1016/j.jogc.2019.05.010
000 JOGC 000 2019
1
OBSTETRICS
INTRODUCTION
iabetes is a relatively common medical condition, affecting 3% to 5% of all pregnancies, with the incidence increasing annually as obesity rates continue to climb.1 Diabetes can lead to a multitude of complications for both the fetus and the mother during pregnancy, including altered fetal growth.2
D
Fetal macrosomia, birth weight exceeding 4000 g, is significantly increased in diabetic pregnancies and is associated with higher rates of perinatal morbidity and mortality, including severe fetal asphyxia, shoulder dystocia, and head trauma.2 In Nova Scotia, between 2015 and 2017, macrosomia was seen in 22% of pregnancies complicated by type 1 diabetes, 19% complicated by type 2 diabetes, and 13% complicated by gestational diabetes mellitus (GDM) (unpublished data). As such, in addition to undergoing a detailed anatomical scan at 18−20 weeks, pregnant women with diabetes will have one or more evaluations of fetal growth throughout the third trimester of pregnancy. As pregnancy approaches term, estimated fetal weight (EFW) obtained from ultrasound is used to guide counselling on mode of delivery to help prevent birth trauma.3,4 The accuracy of EFW from ultrasound measures done in the late third trimester has been questioned, and in non-diabetic women, ultrasound examinations performed at 34−36 weeks gestation may be the most accurate in predicting birth weight.5 Given that glycemic control plays a role in regulating the rate of fetal growth in diabetic pregnancies,6 in the interest of informing the optimal timing of ultrasound in the third trimester for birth weight prediction, the goal of the present study was to determine the accuracy of ultrasound by gestational age (GA) in this patient population. MATERIALS AND METHODS
A retrospective cohort study of diabetic women was conducted. Women for whom third trimester ultrasound scanning was done between 1995 and 2010 at the Fetal Assessment and Treatment Centre (FATC), IWK Health Centre in Halifax, Nova Scotia, were included if they had type 1 diabetes, type 2 diabetes, or GDM; a viable singleton intrauterine pregnancy; and at least one ultrasound examination performed at gestational weeks 28−40. Fetuses with known congenital or chromosomal anomalies were excluded. All information was derived through linkage using individual patient hospital identification numbers from two established databases: the Nova Scotia Atlee Perinatal database (NSAPD) and the IWK Viewpoint
2
000 JOGC 000 2019
database. The NSAPD is administered by the Reproductive Care Program and contains information on all deliveries to residents in Nova Scotia of births ≥20 weeks gestation or ≥500 g. Data are entered by trained coders from standard prenatal and hospital forms. Validation studies on the NSAPD have established that the data are accurate and reliable.7 The NSAPD provided detailed information on maternal characteristics including pre-pregnancy diabetes, pregnancy conditions including GDM, and birth characteristics including GA and birth weight. The Viewpoint database includes information on all ultrasound scans done in the IWK FATC. The Joint Data Access Committee of the Reproductive Care Program of Nova Scotia and the IWK Health Centre Research Ethics Board approved the study and the use of the databases (REB#1009809). The population of interest, women with type 1 diabetes, type 2 diabetes, or GDM, was defined using information from the NSAPD. Over the period of this study, population-wide screening was conducted with glucose challenge testing, and a diagnosis of GDM was made with either a serum glucose of ≥11.1 mmol/L on a 1-hour 50-g oral glucose challenge or one or more abnormal values on a 75-g 2-hour oral glucose tolerance test (fasting serum glucose ≥5.3 mmol/L; 1-hour serum glucose ≥10.6 mmol/L; 2-hour serum glucose ≥9.0 mmol/L). The exposure of main interest was the GA at ultrasound. GA was determined by last menstrual period or by first trimester ultrasound if last menstrual period was uncertain. During the period of study, a routine first trimester scan was not standard of care at our health centre. Experienced nurse ultrasonographers (American Registry for Diagnostic Medical Sonography [ARDMS] certified) or maternal-fetal medicine fellowship−trained physicians performed all ultrasound examinations recorded in the Viewpoint database. The ultrasound measurements extracted from this database were head circumference (HC), biparietal diameter (BPD), femur length (FL), and abdominal circumference (AC). These measurements were used to calculate the EFW at the time of ultrasound using the following Hadlock equation8: Antilog ½1:3596 0:00386 ðACÞ ðFLÞ þ 0:0064 ðHCÞ þ 0:00061 ðBPDÞ ðACÞ þ 0424 ðACÞ þ 0:174 ðFLÞ: We then used the equation proposed by Mongelli to predict the outcome of interest for this study, actual birth weight9:
Optimal Timing of Prenatal Ultrasound in Predicting Birth Weight in Diabetic Pregnancies
Estimated fetal birth weight by ultrasound Median fetal weight at GA of ultrasound Predicted birth weight at delivery ¼ Median fetal weight at GA at delivery Median fetal weight by GA and sex were obtained from a Canadian reference population.10 Actual birth weight was obtained from the NSAPD. Statistical Analysis
The study sample was described using frequencies for categorical characteristics or means with standard deviations (SDs) for continuous characteristics. Analyses conducted to examine the association between GA at ultrasound and accuracy of predicted birth weight considered accuracy in two forms: a continuous variable representing the difference between predicted and actual birth weight; and a categorical variable representing whether predicted birth weight was within 250 g of actual birth weight, underestimated by more than 250 g, or overestimated by more than 250 g. For the analyses with accuracy as a continuous variable, after establishing that data were approximately normally distributed, linear regression was used to determine whether the mean difference between predicted and actual birth weight varied across categories of GA at ultrasound (28−31, 32−35, 36−37, and 38−40 weeks). Multiple linear regression modelling was used to determine whether the association between GA at ultrasound and accuracy was independent of other factors. These factors included any characteristic that was related to mean accuracy with P < 0.10. The following characteristics were considered: maternal age, maternal pre-pregnancy body mass index, parity, smoking, hypertension (pre-existing or gestational), renal disease, previous spontaneous abortion, type of diabetes, and time between ultrasound examination and birth.
RESULTS
The characteristics of the 943 women with diabetes included in this study are shown in Table 1. Of these women, 121 (12.8%) had type 1 diabetes, 111 (11.7%) had type 2 Table 1. Maternal and Neonatal Characteristics N (%) or mean §SD
Characteristic Maternal characteristics Maternal age (years)
31.6 (5.1)
<30
352 (37.3)
30 to <40
543 (57.6)
≥40
48 (5.1)
Type of diabetes Type 1
121 (12.8)
Type 2
111 (11.7)
GDM
711 (75.4)
Pre-pregnancy BMI (kg/m2) 18.5−24.9
210 (22.8)
25−29.9
211 (22.9)
30−34.9
195 (21.2)
35−39.9
157 (17.0)
40 +
149 (16.2)
Maternal smoking
206 (20.3)
Alcohol or drug abuse
62 (6.5)
Hypertension (pre-existing/gestational)
179 (18.9)
Pre-existing heart disease
11 (1.2)
Pre-existing renal disease
22 (2.3)
Pre-existing GI disease
11 (1.2)
Number of ultrasound scans per woman in third trimester 1
815
2
96
3
26
4
4
5
2
For the analyses with accuracy as a categorical variable, logistic regression was used to estimate odds ratios (ORs) with 95% confidence intervals (CIs) for the association between GA at ultrasound (28−31, 32−35, 36−37, and ≥38 weeks) and the odds of underpredicting or overpredicting birth weight by greater than 250 g. Multiple logistic regression was performed to adjust for any covariate that was related to the odds of inaccurate estimation with P < 0.10.
Neonatal characteristics
Proportion LGA newborns
315 (33.4)
Because many women had predicted birth weight from more than one ultrasound scan, all regression analyses were run with generalized estimating equations using the Genmod procedure in SAS software version 9.2 (SAS Institute, Cary, NC).
Birth weight >4000 g
219 (23.2)
Birth weight >4500 g
55 (5.8)
Gestational age at delivery, weeks Delivered at <34 weeks
38.6 § 1.8 21 (2.2)
Delivered at 34 −36 weeks
106 (11.2)
Delivered at ≥370 weeks
816 (86.5)
0
6
Male infant Mean birth weight, g
450 (47.8) 3545 § 652
BMI: body mass index; GDM: gestational diabetes mellitus; GI: gastrointestinal; LGA: large for gestational age.
000 JOGC 000 2019
3
OBSTETRICS
Figure. Scatterplot showing the difference between predicted and actual birth weight by gestational age in weeks. The solid line shows the linear regression line, and the dotted lines are the 95% confidence limits for the individual predicted values. The regression coefficient from adjusted model is 21.5 (95% confidence limits39.1 to 3.9; P = 0.0165).
U/S: ultrasound.
diabetes, and 711 (75.4%) had GDM. The mean time interval between ultrasound examination and delivery was 4.3 (SD 2.8) weeks. The mean birth weight was 3545 (SD 652) g, and the mean difference between predicted birth weight and actual birth weight was 175 (SD 512) g. The Figure shows the scatter plot of GA versus accuracy of ultrasound. The graph shows a trend of the difference between predicted and actual birth weight decreasing as GA at ultrasound increased (P = 0.0165; regression coefficient from the adjusted model is 21.5 [95% CI 39.1 3.9]). Table 2 presents the mean difference between predicted and actual birth weight by GA at ultrasound. Ultrasound
scans performed at 36 weeks or greater were more accurate at predicting birth weight than those performed at an earlier GA. The difference between actual and predicted birth weight was smaller for ultrasound scans performed at 380 weeks or greater than for scans performed at 280−316 weeks (P = 0.002) and at 320−356 weeks (P = 0.001). There was no statistically significant difference between those scans performed at 360−376 weeks compared with 38 weeks or greater (P = 0.08). At 38 weeks or greater, the accuracy of the ultrasound examination was 30 g (95% CI 109 to 48). After adjusting for maternal age, pre-pregnancy body mass index, maternal smoking, type of diabetes, and time between ultrasound examination and delivery, accuracy improved as GA at ultrasound increased (P = 0.005). The ORs for the association between week of gestation at ultrasound examination and underpredicting or overpredicting birth weight by greater than 250 g are shown in Table 3. The odds of underpredicting or overpredicting birth weight by more than 250 g were not significantly affected by the timing of the ultrasound scan. Although not statistically significant, the odds of predicting birth weight to within 250 g of actual birth weight were increased when based on ultrasound examinations completed at ≥38 weeks relative to 280−316 weeks (adjusted OR 1.24; 95% CI 0.62−2.48).
DISCUSSION
In this study, we evaluated the accuracy of third trimester ultrasound examinations in predicting actual birth weight in pregnancies complicated by diabetes. Ultrasound scans at term were the most accurate in predicting actual birth weight compared with earlier GAs. The difference between predicted and actual birth weight decreased as the timing of prenatal ultrasound approached term.
Table 2. Mean difference between predicted and actual birth weight by gestational age (GA) at ultrasound Difference between predicted and actual birth weight, g; Mean (95% CI) Number of ultrasound scans
Unadjusted
Adjusteda
280−316
204
322 (234−410)
247 (147−347)
6
32 −35
409
201 (159−243)
193 (136−251)
360−376
259
63 (13−111)
107 (35−178)
≥38
71
30 (109 to 48)
28 (71 to 126)
GA at ultrasound, weeks
0
a
Means estimated from a linear regression model with generalized estimating equations to account for the repeated ultrasound measures within women and adjustment for maternal age, smoking in pregnancy, pre-pregnancy body mass index, type of diabetes, and time between ultrasound examination and delivery.
4
000 JOGC 000 2019
Optimal Timing of Prenatal Ultrasound in Predicting Birth Weight in Diabetic Pregnancies
Table 3. Odds ratios for the association between gestational age (GA) at ultrasound and underpredicting or overpredicting birth weight Number (row %) GA at ultrasound (weeks)
Unadjusted OR (95% CI)
Adjusted OR (95% CI)a
Underpredicting
§ 250 g
Overpredicting
Underpredicting
Overpredicting
Underpredicting
Overpredicting
280−316
34 (16.2)
56 (26.7)
120 (57.1)
0.81 (0.42−1.57)
2.97 (1.64−5.37)
0.69 (0.29−1.65)
1.62 (0.75−3.48)
320−356
69 (14.4)
185 (38.7)
224 (46.9)
0.50 (0.28−0.88)
1.68 (0.98−2.88)
0.45 (0.23−0.87)
1.23 (0.67−2.24)
360−376
66 (20.2)
154 (47.2)
106 (32.5)
0.57 (0.32−1.02)
0.95 (0.55−1.67)
0.57 (0.31−1.05)
0.86 (0.49−1.51)
≥38
27 (30.3)
36 (40.4)
26 (29.2)
Reference
Reference
Reference
Reference
a
Odds ratios estimated from logistic regression model with generalized estimating equations to account for the repeated ultrasound measures within women and adjustment for maternal age, smoking in pregnancy, pre-pregnancy body mass index, type of diabetes, and time between ultrasound examination and delivery.
Similar to our findings, Cohen et al. demonstrated that ultrasound examinations performed within 4 days of delivery were more accurate in predicting birth weight compared with examinations done more than 4 days before delivery.11 A retrospective multinational cohort study evaluating serial growth scans in non-diabetic women showed that ultrasound examinations performed around 37 weeks gestation had the best accuracy.12 These findings support our results of improved accuracy at term. The impracticality of performing ultrasound scans immediately before delivery lets us have more a realistic time frame on the basis of the results of our study. Because this was a retrospective study, we were limited by the frequency at which women had third trimester growth ultrasound scans. In this study population, only 32 women had more than two third trimester scans. We elected to use the Mongelli equation in this study to overcome this limitation. This equation is a validated method to adjust for lapse times between GA when ultrasound examination is performed and GA at delivery.9 The optimal time point in the third trimester for ultrasound accuracy has remained controversial in the literature, and several other studies have refuted our findings. Two studies using the Mongelli equation, one evaluating nondiabetic pregnancies and another limited to diabetic pregnancies, showed ultrasound scans performed between 340 and 366 weeks gestation to be the most accurate in predicting birth weight.3,5 A recent prospective cohort study showed that ultrasound scans performed earlier in the third trimester were most accurate in predicting small for GA fetuses, whereas scans performed at 36 weeks were most accurate in predicting large for GA fetuses.13 In our study, the average birth weight was 3560 g, and 29% of our cohort had a birth weight greater than 4000 g. Because our cohort consisted of predominately large fetuses, this may explain why we observed improved accuracy of ultrasound examination performed closer to term.
The inaccuracy of ultrasound scans performed at earlier GA in predicting accurate birth weight may be influenced by maternal glycemic control. The Mongelli equation is based on the assumption that fetal growth remains on a growth trajectory defined by weight for GA and sex in a reference population during the third trimester.3,9 Poor glycemic control is associated with an increased rate of fetal growth, thus leading to underestimation of actual birth weight.14 Or, just the opposite, obtaining optimal glycemic control through the third trimester may slow the rate of fetal growth and result in overestimation of birth weight.15 Unfortunately, one of the limitations of our study was that markers of glycemic control were not captured in any of the databases. It is possible that glycemic control affected the ability to predict fetal birth weight accurately with the Mongelli equation with ultrasound measures taken earlier in the third trimester. In addition, over the 15-year period of this cohort, the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) trial was published, which may have influenced management of diabetes in pregnancy.6 Newer novel markers have been studied, including humeral soft tissue thickness, subcutaneous tissue−to-FL ratio, three-dimensional limb volumes, and cheek-to-cheek ratio, to improve the accuracy of ultrasound birth weight prediction in diabetes-affected pregnancies.16,17 Given the retrospective nature of this study, we were not able to incorporate or assess these markers, and it remains an area for future research to determine whether these markers improve the accuracy of third trimester ultrasound examinations in diabetic pregnancies. CONCLUSION
The rationale for third trimester growth ultrasound examinations in diabetic women is to aid in counselling regarding mode of delivery. Predicting fetal weight is more accurate using ultrasound compared with symphysis fundal height at
000 JOGC 000 2019
5
OBSTETRICS
term.18 Ensuring that ultrasound scans are done at the most accurate time is important; it has been shown that overpredicted fetal weight is associated with increased risk for Caesarean delivery.4 The American College of Obstetricians and Gynecologists recommends offering elective Caesarean section in this population if EFW is greater than 4500 g.19 As demonstrated in this study, the optimal time with highest accuracy in predicting birth weight is at term in diabetic pregnancies. Therefore, we suggest performing an ultrasound scan at term in diabetic women to best counsel these women regarding mode of delivery.
7. Joseph KS, Fahey J, Canadian Perinatal Surveillance System. Validation of perinatal data in the discharge abstract database of the Canadian Institute for Health Information. Chronic Dis Can 2009;29:96–100. 8. Combs C, Rosenn B, Miodovnik M, et al. Sonographic EFW and macrosomia: is there an optimum formula to predict diabetic fetal macrosomia? J Matern Fetal Med 2000;9:55–61. 9. Mongelli M, Gardosi J. Gestation-adjusted projection of estimated fetal weight. Acta Obstet Gynecol Scand 1996;75:28–31. 10. Kramer MS, Platt RW, Wen SW, et al. A new and improved populationbased Canadian reference for birth weight for gestational age. Pediatrics 2001;108:e35. 11. Cohen JM, Hutcheon JA, Kramer MS, et al. Influence of ultrasound-todelivery interval and maternal-fetal characteristics on validity of estimated fetal weight. Ultrasound Obstet Gynecol 2010;35:434–41.
Acknowledgements
This research was supported by a Ross Stewart Smith Memorial fellowship from Dalhousie University and an IWK Health Centre Category A grant.
12. Zhang J, Kim S, Grewal J, et al. Predicting large fetuses at birth: do multiple ultrasound examinations and longitudinal statistical modelling improve prediction? Paediatr Perinat Epidemiol 2013;26:199–207.
REFERENCES
14. Ben-Haroush A, Chen R, Hadar E, et al. Accuracy of a single fetal weight estimation at 29-34 weeks in diabetic pregnancies: can it predict large-forgestational-age infants at term? Am J Obstet Gynecol 2007;197. 497.e1−6.
1. Canadian Diabetes Association. Clinical practice guidelines for the prevention and management of diabetes in Canada. Can J Diabetes 2013;37 (Suppl 1):S1–S212. 2. Diabetes Canada Clinical Practice Guidelines Expert Committee, Feig DS, Berger H, et al. Diabetes and pregnancy. Can J Diabetes 2018;42(Suppl 1):S255–82.
13. Tarca AL, Hernandez-Andrade E, Ahn H, et al. Single and serial fetal biometry to detect preterm and term small and large-for-gestational age neonates: a longitudinal cohort study. PLoS One 2016;11:e0164161.
15. Tward C, Barrett J, Berger H, et al. Does gestational diabetes affect fetal growth and pregnancy outcome in twin pregnancies? Am J Obstet Gynecol 2016;214:e1–8.
3. Best G, Pressman EK. Ultrasonographic prediction of birth weight in diabetic pregnancies. Obstet Gynecol 2002;99:740–4.
16. Maruotti GM, Saccone G, Martinelli P. Third trimester ultrasound softtissue measurements accurately predict macrosomia. J Matern Fetal Neonatal Med 2017;30:972–6.
4. Blackwell SC, Refuerzo J, Chadha R, et al. Overestimation of fetal weight by ultrasound: does it influence the likelihood of cesarean delivery for labor arrest? Am J Obstet Gynecol 2009;200. 340.e1−3.
17. Garabedian C, Vambergue A, Salleron J, et al. Prediction of macrosomia by serial sonographic measurements of fetal soft-tissues and the liver in women with pregestational diabetes. Diabetes Metab 2013;39:511–8.
5. Pressman EK, Bienstock JL, Blakemore KJ, et al. Prediction of birth weight by ultrasound in the third trimester. Obstet Gynecol 2000;95:502–6.
18. Ashrafganjooei T, Naderi T, Eshrati B, et al. Accuracy of ultrasound, clinical and maternal estimates of birth weight in term women. East Mediterr Health J 2010;10:313–7.
6. HAPO Study Cooperative Research Group, Metzger BE, Lowe LP, et al. Hyperglycemica and adverse pregnancy outcomes. N Engl J Med 2008;358:1991–2002.
6
000 JOGC 000 2019
19. Committee on Practice Bulletins—Obstetrics. Practice bulletin no. 180: gestational diabetes mellitus. Obstet Gynecol 2017;130:e17–37.
Optimal Timing of Prenatal Ultrasound in Predicting Birth Weight in Diabetic Pregnancies Christopher M. Nash, MD, MSc;1 Christy Woolcott, PhD;1,2 Colleen O’Connell, PhD;1 B. Anthony Armson, MD, MSc1 1
Department of Obstetrics and Gynaecology, Dalhousie University, Halifax, NS
2
Department of Pediatrics, Dalhousie University, Halifax, NS
Abstract
Résumé
Objective: This study sought to determine the optimal timing of ultrasound in the third trimester to predict birth weight accurately in diabetic women with a singleton pregnancy.
tude visait a de terminer le meilleur moment pour effectuer Objectif : L’e chographie au troisie me trimestre chez les femmes une e tiques qui ont une grossesse monofœtale afin de pre dire avec diabe la naissance. exactitude le poids a
Methods: A retrospective cohort study of all diabetic women with a singleton pregnancy treated in Halifax, Nova Scotia, was performed. Estimated fetal weight was derived from ultrasound measures using the Hadlock2 equation. The Mongelli equation was used to predict birth weight. The association between gestational age at ultrasound and accuracy of predicted birth weight was assessed, with accuracy as a continuous variable representing the difference between predicted and actual birth weight and as a categorical variable (with four gestational age categories) representing whether predicted birth weight was within, over, or under 250 g of actual birth weight Results: The cohort of 943 women comprised 121 (12.8%) with type 1 diabetes, 111 (11.7%) with type 2 diabetes, and 711 (75.4%) with gestational diabetes. Ultrasound scans performed at term were the most accurate in predicting birth weight. At this gestational age, the mean difference between predicted and actual birth weight was 30 g (95% confidence interval 109 to 48). After adjusting for maternal body mass index, age, smoking, type of diabetes, and interval between ultrasound examination and delivery, accuracy improved as gestational age at ultrasound increased (P = 0.005). The odds of underpredicting or overpredicting birth weight were not significantly affected by the timing of the ultrasound examination. Conclusion: Because the predictive accuracy of ultrasound prediction of birth weight improves with gestational age, fetal growth assessment at term is recommended to aid with delivery planning in women with diabetes.
Key Words: Estimated fetal weight, diabetes, diagnostic accuracy, ultrasound, birth weight Corresponding author: Dr. Christopher Nash, Department of Obstetrics and Gynaecology, Dalhousie University, Halifax, NS. [email protected] Competing interests: See Acknowledgements. Each author has indicated that they meet the journal’s requirements for authorship.
alise une e tude de cohorte re trospective de toutes Méthodologie : On a re tiques qui connaissaient une grossesse monofœtale les femmes diabe taient traite es a Halifax, en Nouvelle-Ecosse. et qui e Le poids fœtal ae te calcule a partir des mesures e chographiques en utilisant estime pre dire le la formule de Hadlock 2. La formule de Mongelli a servi a la naissance. L’e tude a e value le lien entre l’a ^ ge gestationnel poids a chographie et l’exactitude du poids a la naissance au moment de l’e dit, ou la l’exactitude est une variable continue qui correspond a pre rence entre les poids a la naissance pre dit et re el, ainsi qu’une diffe gories d’a ^ge gestationnel) qui variable nominale (avec quatre cate termine si le poids a la naissance pre dit correspond au poids a la de el ou comporte une diffe rence de plus ou moins 250 g. naissance re Résultats : La cohorte de 943 femmes se composait de 121 femmes te de type 1, 111 femmes (11,7%) (12,8%) atteintes du diabe te de type 2 et 711 femmes (75,4%) atteintes du atteintes du diabe te gestationnel. Les e chographies effectue es a terme ont diabe cet a dit le poids a la naissance avec le plus de pre cision. A ^ge pre rence moyenne entre les poids a la naissance gestationnel, la diffe dit et re el est de 30 g (intervalle de confiance de 95% de 109 pre 48). Apre s un ajustement pour tenir compte de l’indice de masse a ^ge, le tabagisme, le type de diabe te et corporelle maternel, l’a chographique et l’accouchement, la l’intervalle entre l’examen e cision s’est ame liore e a mesure que l’a ^ge gestationnel au pre chographie augmentait (P = 0,005). Le moment de moment de l’e chographique n’a pas eu d’incidence notable sur les l’examen e s de sous-estimer ou sure valuer le poids a la naissance. probabilite diction du poids a la Conclusion : Puisque l’exactitude de la pre l’e chographie s’ame liore avec l’augmentation de l’a ^ge naissance a d’e valuer la croissance fœtale a gestationnel, il est recommande planifier l’accouchement des femmes atteintes terme pour aider a te. du diabe © 2019 The Society of Obstetricians and Gynaecologists of Canada/La Société des obstétriciens et gynécologues du Canada. Published by Elsevier Inc. All rights reserved.
Received on February 15, 2019
J Obstet Gynaecol Can 2019;000(000):1−6
Accepted on May 6, 2019
https://doi.org/10.1016/j.jogc.2019.05.010
000 JOGC 000 2019
1
OBSTETRICS
INTRODUCTION
iabetes is a relatively common medical condition, affecting 3% to 5% of all pregnancies, with the incidence increasing annually as obesity rates continue to climb.1 Diabetes can lead to a multitude of complications for both the fetus and the mother during pregnancy, including altered fetal growth.2
D
Fetal macrosomia, birth weight exceeding 4000 g, is significantly increased in diabetic pregnancies and is associated with higher rates of perinatal morbidity and mortality, including severe fetal asphyxia, shoulder dystocia, and head trauma.2 In Nova Scotia, between 2015 and 2017, macrosomia was seen in 22% of pregnancies complicated by type 1 diabetes, 19% complicated by type 2 diabetes, and 13% complicated by gestational diabetes mellitus (GDM) (unpublished data). As such, in addition to undergoing a detailed anatomical scan at 18−20 weeks, pregnant women with diabetes will have one or more evaluations of fetal growth throughout the third trimester of pregnancy. As pregnancy approaches term, estimated fetal weight (EFW) obtained from ultrasound is used to guide counselling on mode of delivery to help prevent birth trauma.3,4 The accuracy of EFW from ultrasound measures done in the late third trimester has been questioned, and in non-diabetic women, ultrasound examinations performed at 34−36 weeks gestation may be the most accurate in predicting birth weight.5 Given that glycemic control plays a role in regulating the rate of fetal growth in diabetic pregnancies,6 in the interest of informing the optimal timing of ultrasound in the third trimester for birth weight prediction, the goal of the present study was to determine the accuracy of ultrasound by gestational age (GA) in this patient population. MATERIALS AND METHODS
A retrospective cohort study of diabetic women was conducted. Women for whom third trimester ultrasound scanning was done between 1995 and 2010 at the Fetal Assessment and Treatment Centre (FATC), IWK Health Centre in Halifax, Nova Scotia, were included if they had type 1 diabetes, type 2 diabetes, or GDM; a viable singleton intrauterine pregnancy; and at least one ultrasound examination performed at gestational weeks 28−40. Fetuses with known congenital or chromosomal anomalies were excluded. All information was derived through linkage using individual patient hospital identification numbers from two established databases: the Nova Scotia Atlee Perinatal database (NSAPD) and the IWK Viewpoint
2
000 JOGC 000 2019
database. The NSAPD is administered by the Reproductive Care Program and contains information on all deliveries to residents in Nova Scotia of births ≥20 weeks gestation or ≥500 g. Data are entered by trained coders from standard prenatal and hospital forms. Validation studies on the NSAPD have established that the data are accurate and reliable.7 The NSAPD provided detailed information on maternal characteristics including pre-pregnancy diabetes, pregnancy conditions including GDM, and birth characteristics including GA and birth weight. The Viewpoint database includes information on all ultrasound scans done in the IWK FATC. The Joint Data Access Committee of the Reproductive Care Program of Nova Scotia and the IWK Health Centre Research Ethics Board approved the study and the use of the databases (REB#1009809). The population of interest, women with type 1 diabetes, type 2 diabetes, or GDM, was defined using information from the NSAPD. Over the period of this study, population-wide screening was conducted with glucose challenge testing, and a diagnosis of GDM was made with either a serum glucose of ≥11.1 mmol/L on a 1-hour 50-g oral glucose challenge or one or more abnormal values on a 75-g 2-hour oral glucose tolerance test (fasting serum glucose ≥5.3 mmol/L; 1-hour serum glucose ≥10.6 mmol/L; 2-hour serum glucose ≥9.0 mmol/L). The exposure of main interest was the GA at ultrasound. GA was determined by last menstrual period or by first trimester ultrasound if last menstrual period was uncertain. During the period of study, a routine first trimester scan was not standard of care at our health centre. Experienced nurse ultrasonographers (American Registry for Diagnostic Medical Sonography [ARDMS] certified) or maternal-fetal medicine fellowship−trained physicians performed all ultrasound examinations recorded in the Viewpoint database. The ultrasound measurements extracted from this database were head circumference (HC), biparietal diameter (BPD), femur length (FL), and abdominal circumference (AC). These measurements were used to calculate the EFW at the time of ultrasound using the following Hadlock equation8: Antilog ½1:3596 0:00386 ðACÞ ðFLÞ þ 0:0064 ðHCÞ þ 0:00061 ðBPDÞ ðACÞ þ 0424 ðACÞ þ 0:174 ðFLÞ: We then used the equation proposed by Mongelli to predict the outcome of interest for this study, actual birth weight9:
Optimal Timing of Prenatal Ultrasound in Predicting Birth Weight in Diabetic Pregnancies
Estimated fetal birth weight by ultrasound Median fetal weight at GA of ultrasound Predicted birth weight at delivery ¼ Median fetal weight at GA at delivery Median fetal weight by GA and sex were obtained from a Canadian reference population.10 Actual birth weight was obtained from the NSAPD. Statistical Analysis
The study sample was described using frequencies for categorical characteristics or means with standard deviations (SDs) for continuous characteristics. Analyses conducted to examine the association between GA at ultrasound and accuracy of predicted birth weight considered accuracy in two forms: a continuous variable representing the difference between predicted and actual birth weight; and a categorical variable representing whether predicted birth weight was within 250 g of actual birth weight, underestimated by more than 250 g, or overestimated by more than 250 g. For the analyses with accuracy as a continuous variable, after establishing that data were approximately normally distributed, linear regression was used to determine whether the mean difference between predicted and actual birth weight varied across categories of GA at ultrasound (28−31, 32−35, 36−37, and 38−40 weeks). Multiple linear regression modelling was used to determine whether the association between GA at ultrasound and accuracy was independent of other factors. These factors included any characteristic that was related to mean accuracy with P < 0.10. The following characteristics were considered: maternal age, maternal pre-pregnancy body mass index, parity, smoking, hypertension (pre-existing or gestational), renal disease, previous spontaneous abortion, type of diabetes, and time between ultrasound examination and birth.
RESULTS
The characteristics of the 943 women with diabetes included in this study are shown in Table 1. Of these women, 121 (12.8%) had type 1 diabetes, 111 (11.7%) had type 2 Table 1. Maternal and Neonatal Characteristics N (%) or mean §SD
Characteristic Maternal characteristics Maternal age (years)
31.6 (5.1)
<30
352 (37.3)
30 to <40
543 (57.6)
≥40
48 (5.1)
Type of diabetes Type 1
121 (12.8)
Type 2
111 (11.7)
GDM
711 (75.4)
Pre-pregnancy BMI (kg/m2) 18.5−24.9
210 (22.8)
25−29.9
211 (22.9)
30−34.9
195 (21.2)
35−39.9
157 (17.0)
40 +
149 (16.2)
Maternal smoking
206 (20.3)
Alcohol or drug abuse
62 (6.5)
Hypertension (pre-existing/gestational)
179 (18.9)
Pre-existing heart disease
11 (1.2)
Pre-existing renal disease
22 (2.3)
Pre-existing GI disease
11 (1.2)
Number of ultrasound scans per woman in third trimester 1
815
2
96
3
26
4
4
5
2
For the analyses with accuracy as a categorical variable, logistic regression was used to estimate odds ratios (ORs) with 95% confidence intervals (CIs) for the association between GA at ultrasound (28−31, 32−35, 36−37, and ≥38 weeks) and the odds of underpredicting or overpredicting birth weight by greater than 250 g. Multiple logistic regression was performed to adjust for any covariate that was related to the odds of inaccurate estimation with P < 0.10.
Neonatal characteristics
Proportion LGA newborns
315 (33.4)
Because many women had predicted birth weight from more than one ultrasound scan, all regression analyses were run with generalized estimating equations using the Genmod procedure in SAS software version 9.2 (SAS Institute, Cary, NC).
Birth weight >4000 g
219 (23.2)
Birth weight >4500 g
55 (5.8)
Gestational age at delivery, weeks Delivered at <34 weeks
38.6 § 1.8 21 (2.2)
Delivered at 34 −36 weeks
106 (11.2)
Delivered at ≥370 weeks
816 (86.5)
0
6
Male infant Mean birth weight, g
450 (47.8) 3545 § 652
BMI: body mass index; GDM: gestational diabetes mellitus; GI: gastrointestinal; LGA: large for gestational age.
000 JOGC 000 2019
3
OBSTETRICS
Figure. Scatterplot showing the difference between predicted and actual birth weight by gestational age in weeks. The solid line shows the linear regression line, and the dotted lines are the 95% confidence limits for the individual predicted values. The regression coefficient from adjusted model is 21.5 (95% confidence limits39.1 to 3.9; P = 0.0165).
U/S: ultrasound.
diabetes, and 711 (75.4%) had GDM. The mean time interval between ultrasound examination and delivery was 4.3 (SD 2.8) weeks. The mean birth weight was 3545 (SD 652) g, and the mean difference between predicted birth weight and actual birth weight was 175 (SD 512) g. The Figure shows the scatter plot of GA versus accuracy of ultrasound. The graph shows a trend of the difference between predicted and actual birth weight decreasing as GA at ultrasound increased (P = 0.0165; regression coefficient from the adjusted model is 21.5 [95% CI 39.1 3.9]). Table 2 presents the mean difference between predicted and actual birth weight by GA at ultrasound. Ultrasound
scans performed at 36 weeks or greater were more accurate at predicting birth weight than those performed at an earlier GA. The difference between actual and predicted birth weight was smaller for ultrasound scans performed at 380 weeks or greater than for scans performed at 280−316 weeks (P = 0.002) and at 320−356 weeks (P = 0.001). There was no statistically significant difference between those scans performed at 360−376 weeks compared with 38 weeks or greater (P = 0.08). At 38 weeks or greater, the accuracy of the ultrasound examination was 30 g (95% CI 109 to 48). After adjusting for maternal age, pre-pregnancy body mass index, maternal smoking, type of diabetes, and time between ultrasound examination and delivery, accuracy improved as GA at ultrasound increased (P = 0.005). The ORs for the association between week of gestation at ultrasound examination and underpredicting or overpredicting birth weight by greater than 250 g are shown in Table 3. The odds of underpredicting or overpredicting birth weight by more than 250 g were not significantly affected by the timing of the ultrasound scan. Although not statistically significant, the odds of predicting birth weight to within 250 g of actual birth weight were increased when based on ultrasound examinations completed at ≥38 weeks relative to 280−316 weeks (adjusted OR 1.24; 95% CI 0.62−2.48).
DISCUSSION
In this study, we evaluated the accuracy of third trimester ultrasound examinations in predicting actual birth weight in pregnancies complicated by diabetes. Ultrasound scans at term were the most accurate in predicting actual birth weight compared with earlier GAs. The difference between predicted and actual birth weight decreased as the timing of prenatal ultrasound approached term.
Table 2. Mean difference between predicted and actual birth weight by gestational age (GA) at ultrasound Difference between predicted and actual birth weight, g; Mean (95% CI) Number of ultrasound scans
Unadjusted
Adjusteda
280−316
204
322 (234−410)
247 (147−347)
6
32 −35
409
201 (159−243)
193 (136−251)
360−376
259
63 (13−111)
107 (35−178)
≥38
71
30 (109 to 48)
28 (71 to 126)
GA at ultrasound, weeks
0
a
Means estimated from a linear regression model with generalized estimating equations to account for the repeated ultrasound measures within women and adjustment for maternal age, smoking in pregnancy, pre-pregnancy body mass index, type of diabetes, and time between ultrasound examination and delivery.
4
000 JOGC 000 2019
Optimal Timing of Prenatal Ultrasound in Predicting Birth Weight in Diabetic Pregnancies
Table 3. Odds ratios for the association between gestational age (GA) at ultrasound and underpredicting or overpredicting birth weight Number (row %) GA at ultrasound (weeks)
Unadjusted OR (95% CI)
Adjusted OR (95% CI)a
Underpredicting
§ 250 g
Overpredicting
Underpredicting
Overpredicting
Underpredicting
Overpredicting
280−316
34 (16.2)
56 (26.7)
120 (57.1)
0.81 (0.42−1.57)
2.97 (1.64−5.37)
0.69 (0.29−1.65)
1.62 (0.75−3.48)
320−356
69 (14.4)
185 (38.7)
224 (46.9)
0.50 (0.28−0.88)
1.68 (0.98−2.88)
0.45 (0.23−0.87)
1.23 (0.67−2.24)
360−376
66 (20.2)
154 (47.2)
106 (32.5)
0.57 (0.32−1.02)
0.95 (0.55−1.67)
0.57 (0.31−1.05)
0.86 (0.49−1.51)
≥38
27 (30.3)
36 (40.4)
26 (29.2)
Reference
Reference
Reference
Reference
a
Odds ratios estimated from logistic regression model with generalized estimating equations to account for the repeated ultrasound measures within women and adjustment for maternal age, smoking in pregnancy, pre-pregnancy body mass index, type of diabetes, and time between ultrasound examination and delivery.
Similar to our findings, Cohen et al. demonstrated that ultrasound examinations performed within 4 days of delivery were more accurate in predicting birth weight compared with examinations done more than 4 days before delivery.11 A retrospective multinational cohort study evaluating serial growth scans in non-diabetic women showed that ultrasound examinations performed around 37 weeks gestation had the best accuracy.12 These findings support our results of improved accuracy at term. The impracticality of performing ultrasound scans immediately before delivery lets us have more a realistic time frame on the basis of the results of our study. Because this was a retrospective study, we were limited by the frequency at which women had third trimester growth ultrasound scans. In this study population, only 32 women had more than two third trimester scans. We elected to use the Mongelli equation in this study to overcome this limitation. This equation is a validated method to adjust for lapse times between GA when ultrasound examination is performed and GA at delivery.9 The optimal time point in the third trimester for ultrasound accuracy has remained controversial in the literature, and several other studies have refuted our findings. Two studies using the Mongelli equation, one evaluating nondiabetic pregnancies and another limited to diabetic pregnancies, showed ultrasound scans performed between 340 and 366 weeks gestation to be the most accurate in predicting birth weight.3,5 A recent prospective cohort study showed that ultrasound scans performed earlier in the third trimester were most accurate in predicting small for GA fetuses, whereas scans performed at 36 weeks were most accurate in predicting large for GA fetuses.13 In our study, the average birth weight was 3560 g, and 29% of our cohort had a birth weight greater than 4000 g. Because our cohort consisted of predominately large fetuses, this may explain why we observed improved accuracy of ultrasound examination performed closer to term.
The inaccuracy of ultrasound scans performed at earlier GA in predicting accurate birth weight may be influenced by maternal glycemic control. The Mongelli equation is based on the assumption that fetal growth remains on a growth trajectory defined by weight for GA and sex in a reference population during the third trimester.3,9 Poor glycemic control is associated with an increased rate of fetal growth, thus leading to underestimation of actual birth weight.14 Or, just the opposite, obtaining optimal glycemic control through the third trimester may slow the rate of fetal growth and result in overestimation of birth weight.15 Unfortunately, one of the limitations of our study was that markers of glycemic control were not captured in any of the databases. It is possible that glycemic control affected the ability to predict fetal birth weight accurately with the Mongelli equation with ultrasound measures taken earlier in the third trimester. In addition, over the 15-year period of this cohort, the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) trial was published, which may have influenced management of diabetes in pregnancy.6 Newer novel markers have been studied, including humeral soft tissue thickness, subcutaneous tissue−to-FL ratio, three-dimensional limb volumes, and cheek-to-cheek ratio, to improve the accuracy of ultrasound birth weight prediction in diabetes-affected pregnancies.16,17 Given the retrospective nature of this study, we were not able to incorporate or assess these markers, and it remains an area for future research to determine whether these markers improve the accuracy of third trimester ultrasound examinations in diabetic pregnancies. CONCLUSION
The rationale for third trimester growth ultrasound examinations in diabetic women is to aid in counselling regarding mode of delivery. Predicting fetal weight is more accurate using ultrasound compared with symphysis fundal height at
000 JOGC 000 2019
5
OBSTETRICS
term.18 Ensuring that ultrasound scans are done at the most accurate time is important; it has been shown that overpredicted fetal weight is associated with increased risk for Caesarean delivery.4 The American College of Obstetricians and Gynecologists recommends offering elective Caesarean section in this population if EFW is greater than 4500 g.19 As demonstrated in this study, the optimal time with highest accuracy in predicting birth weight is at term in diabetic pregnancies. Therefore, we suggest performing an ultrasound scan at term in diabetic women to best counsel these women regarding mode of delivery.
7. Joseph KS, Fahey J, Canadian Perinatal Surveillance System. Validation of perinatal data in the discharge abstract database of the Canadian Institute for Health Information. Chronic Dis Can 2009;29:96–100. 8. Combs C, Rosenn B, Miodovnik M, et al. Sonographic EFW and macrosomia: is there an optimum formula to predict diabetic fetal macrosomia? J Matern Fetal Med 2000;9:55–61. 9. Mongelli M, Gardosi J. Gestation-adjusted projection of estimated fetal weight. Acta Obstet Gynecol Scand 1996;75:28–31. 10. Kramer MS, Platt RW, Wen SW, et al. A new and improved populationbased Canadian reference for birth weight for gestational age. Pediatrics 2001;108:e35. 11. Cohen JM, Hutcheon JA, Kramer MS, et al. Influence of ultrasound-todelivery interval and maternal-fetal characteristics on validity of estimated fetal weight. Ultrasound Obstet Gynecol 2010;35:434–41.
Acknowledgements
This research was supported by a Ross Stewart Smith Memorial fellowship from Dalhousie University and an IWK Health Centre Category A grant.
12. Zhang J, Kim S, Grewal J, et al. Predicting large fetuses at birth: do multiple ultrasound examinations and longitudinal statistical modelling improve prediction? Paediatr Perinat Epidemiol 2013;26:199–207.
REFERENCES
14. Ben-Haroush A, Chen R, Hadar E, et al. Accuracy of a single fetal weight estimation at 29-34 weeks in diabetic pregnancies: can it predict large-forgestational-age infants at term? Am J Obstet Gynecol 2007;197. 497.e1−6.
1. Canadian Diabetes Association. Clinical practice guidelines for the prevention and management of diabetes in Canada. Can J Diabetes 2013;37 (Suppl 1):S1–S212. 2. Diabetes Canada Clinical Practice Guidelines Expert Committee, Feig DS, Berger H, et al. Diabetes and pregnancy. Can J Diabetes 2018;42(Suppl 1):S255–82.
13. Tarca AL, Hernandez-Andrade E, Ahn H, et al. Single and serial fetal biometry to detect preterm and term small and large-for-gestational age neonates: a longitudinal cohort study. PLoS One 2016;11:e0164161.
15. Tward C, Barrett J, Berger H, et al. Does gestational diabetes affect fetal growth and pregnancy outcome in twin pregnancies? Am J Obstet Gynecol 2016;214:e1–8.
3. Best G, Pressman EK. Ultrasonographic prediction of birth weight in diabetic pregnancies. Obstet Gynecol 2002;99:740–4.
16. Maruotti GM, Saccone G, Martinelli P. Third trimester ultrasound softtissue measurements accurately predict macrosomia. J Matern Fetal Neonatal Med 2017;30:972–6.
4. Blackwell SC, Refuerzo J, Chadha R, et al. Overestimation of fetal weight by ultrasound: does it influence the likelihood of cesarean delivery for labor arrest? Am J Obstet Gynecol 2009;200. 340.e1−3.
17. Garabedian C, Vambergue A, Salleron J, et al. Prediction of macrosomia by serial sonographic measurements of fetal soft-tissues and the liver in women with pregestational diabetes. Diabetes Metab 2013;39:511–8.
5. Pressman EK, Bienstock JL, Blakemore KJ, et al. Prediction of birth weight by ultrasound in the third trimester. Obstet Gynecol 2000;95:502–6.
18. Ashrafganjooei T, Naderi T, Eshrati B, et al. Accuracy of ultrasound, clinical and maternal estimates of birth weight in term women. East Mediterr Health J 2010;10:313–7.
6. HAPO Study Cooperative Research Group, Metzger BE, Lowe LP, et al. Hyperglycemica and adverse pregnancy outcomes. N Engl J Med 2008;358:1991–2002.
6
000 JOGC 000 2019
19. Committee on Practice Bulletins—Obstetrics. Practice bulletin no. 180: gestational diabetes mellitus. Obstet Gynecol 2017;130:e17–37.