ORIGINAL ARTICLE

Obstetric Management in Gestational

DEBORAH L. CONWAY, MD determining the timing and route of de- livery; ● The optimal modality to predict the ptimizing outcomes for women diately any reduction in the perception presence of fetal macrosomia and ex- with mellitus of fetal movements. cessive/disproportionate fetal growth O (GDM) and their requires ● Non-stress testing should be “consid- and the occurrence of shoulder dysto- not only careful metabolic management, ered” after 32 weeks’ gestation in cia and its resulting birth trauma. but also appropriately applied fetal sur- women on insulin and “at or near” term veillance techniques and thoughtful se- in women requiring only dietary man- ANTENATAL FETAL lection of the most advantageous timing agement. SURVEILLANCE — Despite the lack and route of delivery. Whenever possible, ● Biophysical profile testing and Doppler of prospective data in this area, most au- these clinical decisions should be based velocimetry to assess umbilical blood thorities agree that women with GDM on the highest level of evidence available flow “may be considered” in cases of treated with insulin or glyburide, those in and should weigh the likelihood and se- excessive or poor fetal growth, or when poor metabolic control regardless of treat- riousness of both maternal and fetal/ there are comorbid conditions, such as ment modality, and those with comorbid neonatal morbidity. In areas where high- preeclampsia. conditions (such as fetal growth abnor- level evidence is lacking, resources ● Ultrasound should be used to detect fe- malities or hypertension) should undergo should be channeled to designing and im- tal anomalies in women with GDM di- fetal surveillance in the form of non-stress plementing clinical studies to get at good agnosed in the first trimester or with testing, contraction stress testing, or bio- answers. In this review, we examine what fasting glucose levels Ͼ120 mg/dl. physical profile assessments (1,2). Using new information exists in the area of ob- ● to determine fetal lung treatment with insulin as a marker for in- stetric care of women with GDM since the maturity in preparation for delivery is creased fetal risk makes sense, given the time of the Fourth International Work- not necessary in well-dated pregnan- fact that it is these women, not the ones shop-Conference in 1997 and highlight cies after 38 weeks’ gestation. easily controlled with diet, who are more areas where there remains a need for likely to have unrecognized type 2 diabe- sound evidence on which to base practice Timing and route of delivery: tes, a known risk factor for third trimester guidelines. ● The presence of GDM is not by itself an (3). The summary statement from the indication for cesarean delivery. It is unlikely that we will see a ran- 1997 Workshop-Conference remarked ● GDM is not an indication for delivery domized trial specifically addressing the that “the lack of data from controlled clin- before 38 weeks’ gestation in the ab- issue of whether or not women with diet- ical studies on which management rec- sence of evidence of fetal compromise. controlled GDM benefit from additional ommendations can be based was a assessment of fetal well-being beyond prominent theme of discussion regarding daily fetal movement counts. The primary The consensus group lacked suffi- antepartum management of GDM” (1). In reason for this is that the outcomes of in- cient data to draw definitive conclusions the end, consensus was reached in the fol- terest for such a trial ( on the following issues: lowing areas of obstetric management: and long-term neurological morbidities such as ) are relatively rare. Fetal surveillance: ● The need for intensified fetal surveil- For example, to detect a doubling of the ● All women with GDM should monitor lance in women with GDM in good stillbirth rate, an estimated sample size of fetal movements during the last 8–10 control on diet alone; 16,000 women was used (4). weeks of and report imme- ● The role of fetal weight estimation in If such evidence is absent, informa-

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● tion might be drawn from existing and ongoing studies of GDM that include an From the Division of Maternal-Fetal Medicine, Department of and Gynecology, University of unmonitored arm or cohort. In one small Texas Health Science Center–San Antonio, San Antonio, Texas. Address correspondence and reprint requests to Deborah Conway, Assistant Professor, Director, Diabetes randomized trial looking at treatment and in Pregnancy Program, Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, intensive monitoring versus no treatment University of Texas Health Science Center–San Antonio, 7703 Floyd Curl Dr., San Antonio, TX 78229. and no formal fetal surveillance in women E-mail: [email protected]. with GDM, no occurred in the Received for publication 4 April 2006 and accepted in revised form 2 June 2006. 150 women with GDM who had routine This article is based on a presentation at a symposium. The symposium and the publication of this article were made possible by an unrestricted educational grant from LifeScan, Inc., a Johnson & Johnson company. care and no antepartum fetal monitoring Abbreviations: AD-BPD, abdominal diameter–biparietal diameter; EFW, estimated fetal weight; GDM, (5). Casey et al. (6) reported on various gestational diabetes mellitus; HAPO, Hyperglycemia and Adverse Pregnancy Outcome; MFMU, Maternal- outcomes of 874 women with diet- Fetal Medicine Units Network. controlled GDM compared with a large A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion factors for many substances. nondiabetic cohort. Women were classi- DOI: 10.2337/dc07-s212 fied as having diet-controlled GDM if © 2007 by the American Diabetes Association. their fasting glucose on the diagnostic 3-h

DIABETES CARE, VOLUME 30, SUPPLEMENT 2, JULY 2007 S175 Obstetric management in GDM glucose tolerance test was Ͻ105 mg/dl. and comparable to a nondiabetic control tionally charged facets as devastating neo- Fetal surveillance was not performed in cohort, but a relatively small proportion natal injury, avoidance of unnecessary these women unless another indication of them underwent fetal testing, it could maternal harm, and medicolegal liability. for doing so, such as preeclampsia, was be concluded that antenatal fetal testing To compound the problem, the tools and found. Information regarding the propor- does provide additional benefit to women methods we have at our disposal to pre- tion of the diabetic cohort who did un- with “mild” GDM (i.e., absence of fasting dict the maternal-fetal pairs at highest or dergo fetal surveillance was not provided. hyperglycemia). On the other hand, if lowest risk of adverse outcome lack preci- However, the perinatal mortality rate was treatment of mild GDM results in lower sion, while the personal and professional identical between the two groups at mortality and morbidity, but not to the costs of making an incorrect clinical deci- 6/1,000, including comparable stillbirth level found in the nondiabetic women, it sion remain high. Although some progress rates (5/1,000 in the women with diet- may be that fetal surveillance before 40 has been made since the Fourth Interna- controlled GDM and 4/1,000 in the non- weeks has a place in identifying the preg- tional Workshop-Conference in 1997, diabetic women) (6). Notably, this cohort of nancies at highest risk and preventing ad- much work remains to be done. women with diet-treated GDM may or may verse outcome. At its core, this argument boils down not have been “diet-controlled,” given the The HAPO study is enrolling 25,000 to a diligent and thoughtful weighing of high rate of large-for-gestational-age and women in 10 countries who will undergo maternal and fetal risks: the chance of se- macrosomic infants (35 and 23%, respec- 2-h 75-g glucose tolerance tests. As long vere damage to the mother with GDM tively), but nonetheless experienced still- as the fasting glucose level is Ͻ105 mg/dl from cesarean delivery versus the chance birth rates comparable to a nondiabetic and the 2-h postload value is below 200 of severe damage to her from a population without universal antenatal mg/dl, the results of the glucose tolerance shoulder dystocia event at vaginal deliv- fetal surveillance. Thus, the severity of the test will not be revealed to care providers, ery. Fortunately, both occurrences are disease process as evidenced by the glu- and all women in the cohort will be fol- relatively rare. The risk of brachial plexus cose tolerance test results (i.e., the pres- lowed prospectively for various preg- injury (at least transient) when a macro- ence or absence of fasting hyperglycemia) nancy outcomes. The sample size was somic infant (Ͼ4,000 g) is delivered vag- may be a better indicator of fetal risk and planned to provide sufficient numbers inally by a diabetic woman is ϳ2–5% the need for antepartum fetal testing than across the spectrum of glucose values (10–12). Compiling data from several re- the treatment modality. Women with with the intent to identify thresholds use- ports, Rouse and Owen (16) estimated GDM who have fasting hyperglycemia are ful for predicting morbidity attributable that the mean probability that a brachial more likely to require insulin to control to GDM. For example, it is estimated that plexus injury will persist is 6.7%. There- their glucose levels, but the decision to ϳ400 women in the cohort will have a fore, out of 10,000 vaginal deliveries of add such treatment is more subjective fasting glucose value between 100 and macrosomic infants, ϳ13–33 will result than the glucose tolerance test results. 105 mg/dl (8). A large proportion of these in persistent , of Two large multicenter studies are cur- women will not undergo treatment or which roughly three-quarters can expect rently underway to determine the impact tests of fetal well-being. Thus, the HAPO full recovery of shoulder and elbow func- of GDM on obstetric and perinatal out- study, like the MFMU trial, will provide tion when surgery is performed in the first comes: one conducted through the Ma- unprecedented data concerning the need year of life (17). Even with complete bra- ternal-Fetal Medicine Units Network for specialized fetal surveillance in preg- chial plexus palsy, involving the hand and (MFMU) of the National Institute for nancies complicated by relatively mild associated with Horner’s syndrome, Child Health and Human Development glucose intolerance. staged surgical intervention over the first (NICHD) and the multinational Hyper- 3–4 years of life can provide useful hand glycemia and Adverse Pregnancy Out- SHOULDER DYSTOCIA IN function in 76% of cases (18). Nonethe- come (HAPO) study. Both studies will GDM: PITFALLS IN less, avoidance of such an outcome in the provide prospective outcome data on a PREDICTION AND first place is preferable, leading us to con- large number of women with GDM who PREVENTION — It is clear that sider the maternal burden of morbidity do not receive treatment. The MFMU women with GDM are at increased risk from elective prelabor cesarean delivery. GDM trial will involve 950 women with both for delivering an excessively grown It is widely assumed that cesarean de- abnormal glucose tolerance tests, but fast- infant and for having that delivery com- livery results in more maternal morbidity ing glucose Ͻ95 mg/dl, randomized to plicated by shoulder dystocia (9). When and, indeed, mortality than vaginal deliv- treatment or routine care. These groups shoulder dystocia occurs, infants of ery, and some evidence exists for a two- to will be compared to a matched cohort of mothers with diabetes are more likely to fourfold greater risk of in women with normal glucose screening re- incur brachial plexus injury than infants women delivered by cesarean delivery sults. Fetal surveillance in the form of of nondiabetic women (10–12). How- compared with (19). non-stress tests will not occur in the ever, the best strategy for avoiding this However, women with complications treated group until 40 weeks’ gestation, outcome is a controversial topic, usually that increase the risk of maternal death although women in either group may un- centered on the use of cesarean delivery to and serious morbidity, such as severe hy- dergo fetal testing for other routine ob- prevent difficult vaginal birth and thus in- pertensive disease, hemorrhage from pla- stetric indications (7). Thus, this trial will jury to the infant. Although brachial centa previa or abruption, true obstructed give us information on almost 1,000 plexus injury after cesarean delivery has labor, and life-threatening infections, of- women with mild GDM who, by-and- been described (13,14), it is an exceed- ten are delivered by cesarean section, large, will not be monitored with non- ingly rare event (15). Unfortunately, few making it difficult to discern the risk at- stress tests. If outcomes are better in the data currently exist to put an end to the tributable to the operative intervention it- treated group than the untreated group controversy, which involves such emo- self. Conversely, it is difficult to find data

S176 DIABETES CARE, VOLUME 30, SUPPLEMENT 2, JULY 2007 Conway indicating that an elective prelabor cesar- roughly equivalent accuracy, even in GDM has emerged since the Fourth Inter- ean delivery at term is any riskier than macrosomic fetuses (22–24), making it national Workshop-Conference. Yogev et vaginal delivery. Information from the difficult to recommend one method over al. (27) reported their experience with a Washington state birth events records the other based on hard evidence. None- clinical protocol in which 84 women with database from 1990 indicates that theless, obtaining a fetal weight estimate GDM underwent at women delivering a macrosomic infant by ultrasound provides some measure of 38–39 weeks if they were treated with by prelabor cesarean section have a objectivity over clinical estimation and insulin and/or the fetus was above the threefold greater risk of postpartum in- has been shown to be as accurate in obese 90th percentile (but below 4,000 g) by fection, an 11-fold greater risk of women as in lean women (25). ultrasound estimation. The overall cesar- wound complications, but an 80% Some evidence exists that using ultra- ean delivery rate in this cohort was 18%, lower risk of postpartum hemorrhage sound-derived fetal weight estimates to significantly higher than in a cohort of than women experiencing a vaginal de- inform decisions regarding timing and nondiabetic women in spontaneous labor livery of a macrosomic infant. Overall, route of delivery in diabetic women can (9%), but no different than in a group of rates of each were low in result in lowered rates of shoulder dysto- nondiabetic women who underwent elec- both groups (15). However, the cumu- cia. We published our experience with a tive labor induction (14.8%). This clinical lative risk of repeated cesarean deliver- clinical policy of obtaining an ultrasonic policy resulted in a macrosomia rate of ies needs also to be considered and estimated fetal weight (EFW) at 37–38 only 5.7% in this group of women. No factored into clinical decision making. weeks in women with diabetes who were comparison to a similar population man- Besides the well-known risks to the eligible for vaginal delivery and used the aged without this protocol is provided, mother of placenta previa/accreta with a results as follows: when the EFW was and therefore the impact of this practice uterine scar (20), newer data indicate Ͼ4,250 g, cesarean delivery was recom- on outcomes cannot be determined. that prior cesarean delivery increases mended, and when the EFW was above Shoulder dystocia rate is also not re- the risk for stillbirth in subsequent the 90th percentile (and below 4,250 g), ported. (21). labor induction was performed. We com- How might we refine and improve Thus, it appears that avoiding vaginal pared the shoulder dystocia rates among our approach to selecting the maternal- delivery benefits the infant destined to 1,337 women managed under this proto- fetal pairs that would most benefit from suffer shoulder dystocia and brachial col to a historical cohort of 1,227 women avoiding vaginal delivery? More accurate plexus injury, whereas elective prelabor managed without; antenatal management estimation of fetal weight/prediction of cesarean delivery poses relatively minor of diabetes was otherwise similar between birth weight would minimize maternal risk to mothers. The key, then, is to accu- the two time periods. The shoulder dys- morbidity from cesarean sections done in rately identify the maternal-fetal pairs tocia rate in the cohort in whom the EFW error for suspected macrosomia. Alterna- who need such intervention and allow the protocol was used was significantly de- tively, being able to accurately detect the others to labor. However, we currently creased in the overall diabetic population: fetal body asymmetry and/or the fetal- lack the capability to do so with accept- 1.5 vs. 2.4% (OR 0.5, 95% CI 0.3–1.0). pelvic disproportion that might contrib- able precision. The problem is that iden- The largest impact was found in the fe- ute to shoulder dystocia (and perhaps tifying the large fetus is not enough. We tuses at greatest risk: the shoulder dysto- brachial plexus injury) risk would be really want to identify the fetus whose ex- cia rate among macrosomic infants was helpful. Currently, however, little data cessive disproportionate growth will re- 19% before the EFW protocol and 7% us- exist along these lines. Magnetic reso- sult in its negotiating the birth canal to a ing the protocol (OR 0.3, 95% CI 0.1– nance imaging and three-dimensional ul- sufficient degree to prevent arrested la- 1.0). The EFW protocol affected the trasound are promising new modalities bor, but who will then experience a shoul- timing and route of delivery of only that may improve fetal weight estimation der dystocia. Once a shoulder dystocia 10.6% of our diabetic population (6.8% by providing volumetric assessments of occurs, its recognition and management labor induction and 3.8% cesarean deliv- the fetus. Results from various reports are may affect the likelihood of brachial ery), which compares favorably to the 9% summarized in Table 1. In general, both plexus injury. On the other hand, there rate of macrosomia that was found in this three-dimensional ultrasound and mag- may be something we don’t understand cohort. Despite this relatively low rate of netic resonance imaging result in more about the interaction between the mater- intervention, the protocol probably re- accurate fetal weight estimates than two- nal and soft tissues and the fetus sulted in a significant increase in our over- dimensional ultrasound. Most of these that makes a shoulder dystocia more dif- all cesarean delivery rate among diabetic studies are limited in their applicability to ficult to relieve, thus placing the infant at women after its implementation (25.1%, the issue of fetal weight estimation in di- increased risk for injury despite our most up from 21.7% in the earlier time period) abetic women for several reasons: overall careful maneuvers. We currently lack the (26). These data also have the advantage sample sizes are small, few include or ability to get at these complex interactions of being derived from a single center’s have much less focus on a diabetic popu- in a clinically useful way. population, pointing out the important, lation, and fetuses at the extremes of However, the well-intentioned desire but poorly studied, impact of local prac- weight are few in number. In addition, the to avoid birth trauma remains, and thus tice styles, baseline macrosomia and ce- performance of these modalities in rou- we attempt to antenatally detect the large sarean delivery rates, and patient tine clinical use (i.e., outside of a research fetus, who is more likely to suffer a shoul- population characteristics on the cost- setting) has not been evaluated, and their der dystocia and nerve injury (9,11). The benefit balance of cesarean delivery to availability and cost are also potential ob- two most widely available means of esti- prevent brachial plexus injury. stacles. mating fetal weight, clinical assessment Little additional information regard- Although a great deal of work has and ultrasound, have been shown to have ing the timing of delivery in women with been done to sonographically identify the

DIABETES CARE, VOLUME 30, SUPPLEMENT 2, JULY 2007 S177 Obstetric management in GDM

Table 1—Summary of studies of volumetric assessment of fetal weight

Imaging Author modality n Include diabetic women? Findings Comments Uotila et al. (30) MRI 20 Yes, 10 type 1 diabetes MRI (10-mm sagittal slices) “Most” fetuses were LGA; more accurate than 2D US at imaging completed in predicting BW under 1 min Schild et al. (31) 3D US 125 ϩ 65 Unknown 3D US formula (thigh volume, 13 macrosomic infants, upper arm volume, abdominal 22 LGA infants volume, BPD) more accurate than 2D US at predicting BW Tukeva et al. (32) MRI 8 Yes, all MRI measurement (TrueFISP) All had suspected of fetal shoulder width was macrosomia; 7/8 were closely correlated with actual delivered by cesarean shoulder width section Hassibi et al. (33) MRI 35 Unknown MRI (8-mm axial images) Imaging time was 90 s significantly better than 2D US (Hadlock) at predicting BW Lee et al. (34) 3D US 100 Unknown 3D US (fractional thigh volume ϩ AC) significantly better than 2D US at predicting BW 2D, two-dimensional; 3D, three-dimensional; AC, abdominal circumference; BPD, biparietal diameter; BW, birth weight; LGA, large-for-gestational-age; MRI, magnetic resonance imaging; TrueFISP, true fast imaging with steady-state precession; US, ultrasound.

“fat” fetus, with varying degrees of suc- ber. In another study, a cohort of 84 brachial plexus injury, at least shoulder cess, most studies have reported the accu- women with infant birth weight Ͼ4,000 g dystocia. racy of predicting birth of a large infant, was compared with 84 women delivering rather than correlating these findings to nonmacrosomic infants. Of the 65 vaginal SUMMARY — Reviewing the areas of obstetric outcomes such as shoulder dys- deliveries in the macrosomic group, 13% controversy related to the obstetric man- tocia or labor abnormalities. Cohen et al. were complicated by shoulder dystocia. agement of women with GDM, we are un- (28) describe the use of the abdominal The authors found that an abdominal cir- fortunately unable to provide significant diameter–biparietal diameter (AD-BPD) cumference of at least 35 cm, obtained refinement of the recommendations difference to identify a pregnancy at risk within 2 weeks of delivery, had a sensitiv- agreed upon after the Fourth Interna- for shoulder dystocia. In a group of 31 ity of 93% and specificity of 88% for mac- tional Workshop-Conference due to the women with diabetes, all of whom were rosomia (29). They describe positive and lack of properly controlled and powered suspected of carrying a large fetus, they negative predictive values that appeared clinical studies in this area since 1997. In found no difference in maternal charac- good, but were invalid because of the the area of the need for antenatal fetal sur- teristics or birth weight between the de- case-control design of the study. Thus, we veillance in women with milder degrees liveries complicated by severe shoulder are again left with sonographic findings of GDM, we may be able to draw indirect dystocia and those with uncomplicated with good sensitivity, but unknown (and conclusions from ongoing cohort studies vaginal delivery. However, the mean AD- likely poor) positive predictive value, the that will include large numbers of BPD difference was higher in the shoulder component of test accuracy that would be women. In the area of optimal timing and ϭ dystocia group (3.1 vs. 2.6, P 0.05). most helpful in identifying before delivery mode of delivery to avoid fetal injury, The cutoff of 2.6 resulted in 100% sensi- those at highest risk for difficult birth. large well-controlled prospective studies tivity (all cases of shoulder dystocia were To help resolve this issue of the value do not currently exist and are urgently above this cutoff) and 46% specificity. of sonographic or magnetic resonance im- needed. In addition, refinement of fetal The authors describe a 30% positive pre- aging detection of the excessively grown and pelvic imaging techniques to more dictive value in predicting severe shoul- fetus, we require studies with the follow- accurately identify the maternal-fetal der dystocia for an AD-BPD difference of ing characteristics: a large cohort of dia- pairs most likely to benefit from avoiding at least 2.6, but this number is almost cer- betic women; imaging obtained within a vaginal delivery, and the more wide- tainly an overestimate, given the fact that short time frame before delivery and in all spread availability of these technologies, their cohort included only vaginal deliv- women regardless of clinical estimate of may also prove to be of benefit in the ob- eries. It is likely that a substantial propor- fetal size; comprehensive ultrasound stetric management of women with GDM. tion of large fetuses with an AD-BPD measurements, including specialized difference above this threshold will un- measures such as shoulder soft tissue dergo cesarean deliveries for labor abnor- thickness and cheek-to-cheek diameter, References malities; inclusion of these cases in the in all women; and a reasonably high rate 1. Metzger BE, Coustan DR: Summary and denominator of the positive predictive of vaginal delivery from which we may recommendations of the Fourth Interna- value calculation would lower that num- begin to identify better markers for, if not tional Workshop-Conference on Gesta-

S178 DIABETES CARE, VOLUME 30, SUPPLEMENT 2, JULY 2007 Conway

tional Diabetes Mellitus. Diabetes Care 21: plexus injury. Am J Obstet Gynecol 179: 25. Field NT, Piper JM, Langer O: The effect B161–B167, 1998 686–689, 1998 of maternal on the accuracy of fe- 2. American College of Obstetricians and 13. Morrison JC, Sanders JR, Magann EF, tal weight estimation. Obstet Gynecol 86: Gynecologists: ACOG practice bulletin: Wiser WL: The diagnosis and manage- 102–107, 1995 gestational diabetes. Obstet Gynecol 98: ment of dystocia of the shoulder. Surg Gy- 26. Conway DL, Langer O: Elective delivery 525–538, 2001 necol Obstet 175:515–522, 1992 of infants with macrosomia in diabetic 3. Cundy T, Gamble G, Townend K, Henley 14. Bar J, Dvir A, Hod M, Orvieto R, Merlob P, women: reduced shoulder dystocia versus PG, MacPherson P, Roberts AB: Perinatal Neri A: Brachial plexus injury and obstet- increased cesarean deliveries. Am J Obstet mortality in type 2 diabetes mellitus. Dia- rical risk factors. Int J Gynaecol Obstet 73: Gynecol 178:922–925, 1998 bet Med 17:33–39, 2000 21–25, 2001 27. Yogev Y, Ben-Haroush A, Chen R, Glick- 4. Landon MB, Vickers S: Fetal surveillance 15. Gregory KD, Henry OA, Ramicone E, man H, Kaplan B, Hod M: Active induc- in pregnancy complicated by diabetes Chan LS, Platt LD: Maternal and infant tion management of labor for diabetic mellitus: is it necessary? J Matern Fetal complications in high and normal weight pregnancies at term: mode of delivery and Neonatal Med 12:413–416, 2002 infants by method of delivery. Obstet Gy- fetal outcome: a single center experience. 5. Garner P, Okun N, Keely E, Wells G, Per- necol 92:507–513, 1998 kins S, Sylvain J, Belcher J: A randomized 16. Rouse DJ, Owen J: Prophylactic cesarean Eur J Obstet Gynecol 114:166–170, 2004 controlled trial of strict glycemic control delivery for fetal macrosomia diagnosed 28. Cohen B, Penning S, Major C, Ansley D, and tertiary level obstetric care versus by means of ultrasonography: a Faustian Porto M, Garite T: Sonographic predic- routine obstetric care in the management bargain? Am J Obstet Gynecol 181:3 tion of shoulder dystocia in infants of di- of gestational diabetes: a pilot study. Am J 32–338, 1999 abetic mothers. Obstet Gynecol 88:10–13, Obstet Gynecol 177:190–195, 1997 17. Dumont CE, Forin V, Asfazadourian H, 1996 6. Casey BM, Lucas MJ, Mcintire DD, Lev- Romana C: Function of the upper limb 29. Jazayeri A, Heffron JA, Phillips R, Spellacy eno KJ: Pregnancy outcomes in women after surgery for obstetric brachial plexus WN: Macrosomia prediction using ultra- with gestational diabetes compared with palsy. J Bone Joint Surg Br 83:894–900, sound fetal abdominal circumference of the general obstetric population. Obstet 2001 35 centimeters or more. Obstet Gynecol Gynecol 90:869–873, 1997 18. Haerle M, Gilbert A: Management of com- 93:523–526, 1999 7. Landon MB, Thom E, Spong CY, Gabbe plete obstetric brachial plexus lesions. 30. Uotila J, Dastidar P, Heinonen T, Ryymin SG, Leindecker S, Johnson F, Lain K, Mi- J Pediatr Orthop 24:194–200, 2004 P, Punnonen R, Laasonen E: Magnetic res- odovnik M, Carpenter M: A planned ran- 19. Miller JM Jr: Maternal and neonatal mor- onance imaging compared to ultrasonog- domized clinical trial of treatment for bidity and mortality in cesarean section. raphy in fetal weight and volume mild gestational diabetes mellitus. J Ma- Obstet Gynecol Clin North Am 15: estimation in diabetic and normal preg- tern Fetal Neonatal Med 11:226–231, 629–638, 1988 nancy. Acta Obstet Gynecol Scand 79:255– 2002 20. Chazotte C, Cohen WR: Catastrophic 259, 2000 8. HAPO Study Cooperative Research complications of previous cesarean sec- 31. Schild RL, Fimmers R, Hansmann M: Fe- Group: The hyperglycemia and adverse tion. Am J Obstet Gynecol 163:738–742, tal weight estimation by three-dimen- pregnancy outcome (HAPO) study. Int J 1990 sional ultrasound. Ultrasound Obstet Gynaecol Obstet 78:69–77, 2002 21. Smith GC, Pell JP, Dobbie R: Caesarean Gynecol 16:445–452, 2000 9. Langer O, Berkus MD, Huff RW, Samue- section and risk of unexplained stillbirth 32. Tukeva TA, Salmi H, Poutanen VP, Kar- loff A: Shoulder dystocia: should the fetus in subsequent pregnancy. Lancet 362: jalainen PT, Hytinantti T, Paavonen J, weighing greater than or equal to 4000 1779–1784, 2003 Teramo KA, Aronen HJ: Fetal shoulder grams be delivered by cesarean section? 22. Sherman DJ, Arieli S, Tovbin J, Siegel G, measurements by fast and ultrafast MRI Am J Obstet Gynecol 165:831–837, 1991 Caspi E, Bukovsky I: A comparison of 10. Kolderup LB, Laros RK Jr, Musci TJ: Inci- clinical and ultrasonic estimation of fetal techniques. J Magn Reson Imaging 13: dence of persistent birth injury in macro- weight. Obstet Gynecol 91:212–217, 1998 938–942, 2001 somic infants: association with mode of 23. Mehdizadeh A, Alaghehbandan R, Horsan 33. Hassibi S, Farhataziz N, Zaretsky M, delivery. Am J Obstet Gynecol 177:37–41, H: Comparison of clinical versus ultra- McIntire D, Twickler DM: Optimization 1997 sound estimation of fetal weight. Am J of fetal weight estimates using MRI: com- 11. Ecker JL, Greenberg JA, Norwitz ER, Perinatol 17:233–236, 2000 parison of acquisitions. AJR Am J Roentge- Nadel AS, Repke JT: Birth weight as a pre- 24. Chauhan SP, West DJ, Scardo JA, Boyd nol 183:487–492, 2004 dictor of brachial plexus injury. Obstet JM, Joiner J, Hendrix NW: Antepartum 34. Lee W, Deter RL, Ebersole JD, Huang R, Gynecol 89:643–647, 1997 detection of macrosomic fetus: clinical Blanckaert K, Romero R: Birth weight 12. Bryant DR, Leonardi MR, Landwehr JB, versus sonographic, including soft-tissue prediction by three-dimensional ultra- Bottoms SF: Limited usefulness of fetal measurements. Obstet Gynecol 95: sonography: fractional limb volume. J Ul- weight in predicting neonatal brachial 639–642, 2000 trasound Med 20:1283–1292, 2001

DIABETES CARE, VOLUME 30, SUPPLEMENT 2, JULY 2007 S179