| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
by the American Institute of Ultrasound in Medicine J Ultrasound Med 27:39-43 • 0278-4297 Predicting MacrosomiaDepartment of Obstetrics and Gynecology, University of Texas Southwestern Medical Center, Dallas, Texas USA. Address correspondence to Jason A. Pates, MD, Department of Obstetrics and Gynecology, Madigan Army Medical Center, Fitzsimmons Drive, Building 9040, Tacoma, WA 98431 USA. E-mail jason.pates{at}nw.amedd.army.mil
Objective. The purpose of this study was to evaluate the prediction of fetal macrosomia based on ultrasound estimates of fetal weight and amniotic fluid volume combined with clinical risk factors. Methods. A retrospective cohort study of women undergoing indicated obstetric ultrasound examinations within 7 days of delivery was conducted. Results. A total of 3115 women gave birth within 7 days of ultrasound examinations that included an estimated fetal weight (EFW) and an amniotic fluid index (AFI). Clinical risk factors were associated with an 8% positive predictive value for a birth weight of 4000 g or higher. Adding an ultrasound EFW of 4000 g or higher increased the positive predictive value to 62%. Adding an AFI of 20 cm or higher to the clinical risk factors and the ultrasound EFW further increased the positive predictive value to 71%. Conclusions. An ultrasound EFW of 4000 g or higher within 1 week of delivery combined with clinical risk factors and an increased AFI is associated with macrosomia at birth in 71% of cases.
Key Words: amniotic fluid index • macrosomia • prediction • ultrasound Abbreviations: AFI, amniotic fluid index • BMI, body mass index • EFW, estimated fetal weight
The antenatal diagnosis of fetal macrosomia remains problematic despite continuing advances in obstetric ultrasound.1,2 Commonly defined as a birth weight of 4000 g or higher, fetal macrosomia is associated with increased maternal and neonatal morbidity.2 Studies of ultrasound for prediction of fetal size published during the 1990s generally showed no improvement over clinical techniques such as Leopold maneuvers.3–6 More recent data, however, suggest that ultrasound may have an increased positive predictive value for diagnosing macrosomia when the amniotic fluid index (AFI) is also considered. For example, it has been observed that mild to moderate hydramnios has a strong association with neonates who are in the 90th percentile or higher for gestational age.6–9 In case-control studies without ultrasound, clinical findings such as diabetes, parity, the maternal body mass index (BMI), a male fetus, a gestational age of greater than 40 weeks, and race have been identified to be risk factors for fetal macrosomia.10 Such risk factors have been recently combined with ultrasound findings to improve the detection of macrosomia, with resultant positive predictive values of nearly 85%.11,12 However, the high prevalence of macrosomia in these study cohorts may decrease the applicability of such models in obstetric populations with a lower prevalence of macrosomia. We hypothesized that incorporating ultrasound estimates of fetal weight with the aforementioned maternal risk factors plus the AFI in a large group of women with a lower prevalence of macrosomia may better predict fetuses whose birth weights are actually 4000 g or higher and may more accurately reflect the utility of such prediction models.
This retrospective cohort study was approved by the Institutional Review Board at the University of Texas Southwestern Medical Center. Maternal demographic features, prenatal care data, labor characteristics, ultrasound findings, and delivery information were entered into computerized clinical databases. These databases were queried to identify all women with singleton pregnancies who underwent obstetric ultrasound examinations that included an anatomic survey of the fetus and an AFI measured within 1 week of delivery. Fetal malformations were excluded. The ultrasound examinations were performed by certified obstetric sonographers under the supervision of maternal-fetal medicine physicians. The ultrasound images were acquired with a curvilinear 3.5- or 5.0-MHz transducer and either a Sonoline Elegra or a Sonoline Antares system (Siemens Medical Solutions USA, Inc, Mountain View, CA). An anatomic survey and fetal biometric data were obtained during each study, in accordance with established American Institute of Ultrasound in Medicine standards.13 The 4-quadrant AFI was obtained by the method described by Phelan and colleagues.14 The estimated fetal weight (EFW) was derived from the abdominal circumference, biparietal diameter, head circumference, and femur length according to the weight formula developed by Hadlock et al.15 An AFI of 20 cm or higher, corresponding to approximately the 95th percentile at 40 weeks, was used as the AFI threshold for this study.16
The computerized obstetric database was accessed to identify women with maternal and fetal risk factors for macrosomia as described by the American College of Obstetricians and Gynecologists.1 These variables included diabetes, the maternal BMI, multiparity (defined as
Statistical analyses were performed with SAS 9.1 (SAS Institute Inc, Cary, NC). Univariate analyses were performed with the
A total of 3115 women gave birth within 7 days of obstetric ultrasound examinations between August 1997 and February 2006, and 237 (7.6%) of these women had neonates weighing 4000 g or higher (Figure 1  ). The median interval from the ultrasound examination to delivery was 3 days. Approximately 61% of fetuses with ultrasound EFWs of 4000 g or more actually weighed 4000 g or higher at birth (Figure 1). Approximately 57% of the ultrasound examinations were performed to assess fetal size, and 23% were done for complications such as hypertensive disorders (n = 369) and diabetes (n = 344), most of the latter being gestational diabetes. The remainder were done for a large variety of indications, none individually exceeding 2.5% of the study population.
Maternal and fetal clinical characteristics in women with neonates having birth weights of 4000 g or higher and those lower than 4000 g are compared in Table 1 . Every maternal or fetal characteristic except maternal race and fetal sex was a significant risk factor for macrosomia at birth. These clinical risk factors were then integrated with an ultrasound EFW of 4000 g or higher as well as an AFI of 20 cm or higher to determine whether the combination of these clinical and ultrasound findings would modify the prediction of an actual birth weight of 4000 g or higher (Table 2). Women with 1 or more risk factors coupled with an ultrasound EFW of 4000 g or higher and an AFI of 20 cm or higher had a 71% positive predictive value for macrosomia at birth. The positive predictive value remained unchanged when the women were stratified by diabetes status. The most powerful single predictive factor for macrosomia was an ultrasound EFW of 4000 g or higher (positive predictive value, 61%). One or more clinical risk factors and an AFI of 20 cm or higher had much lower positive predictive values (8% and 26%, respectively) when used alone.
In this study of 3115 women who underwent ultrasound examinations within 7 days of delivery, the addition of clinical risk factors and an AFI of 20 cm or higher to ultrasound estimates of a birth weight of 4000 g or higher improved the prediction of macrosomia from 61% to 71%. Demographic features such as maternal parity, the BMI, and diabetes were all confirmed to be risk factors for macrosomia, as previously reported, albeit with low positive predictive values when used alone. Similarly, the AFI by itself improved the accuracy of prediction over clinical risk factors alone, but the predictive value was still quite low (26%). Two recently published studies evaluated the prediction of macrosomia using similar ultrasound findings and patient risk factors.11,12 Mazouni and colleagues11 evaluated 194 women with a prediction nomogram based on parity, ethnicity, the BMI, and the ultrasound EFW, with a positive predictive value of nearly 85% for a birth weight of 4000 g or higher. Similarly, Hackmon et al12 combined the third-trimester AFI with the ultrasound EFW and found that an AFI in the 60th percentile or higher combined with an EFW in the 71st percentile predicted 85% of macrosomic neonates. These reports, however, had study cohorts with prevalence rates of macrosomia of 55% and 33%, respectively, thus limiting their applicability to most obstetric populations, in which the prevalence of macro-somia is around 7%.17 Chauhan and colleagues18 made a similar observation in a review of 20 reports on the ultrasound prediction of macrosomia. They found that the positive predictive values varied greatly from study to study and were highly influenced by maternal risk factors as well as the prevalence of macrosomia in the population studied.18 To our knowledge, our simple combination of readily identified clinical risk factors with the clinical ultrasound EFW and AFI in a population with a relatively low prevalence of macrosomia (7.6%) has not been described previously. Although our results suggest that up to 71% of neonates predicted to be macrosomic were in fact macrosomic, we must emphasize that our results, like those in similar studies, were affected by the circumstances under which ultrasound examinations were performed in this retrospective cohort. Namely, often times, women in our study underwent ultrasound evaluations because they were perceived to be at an increased risk for macrosomia because of diabetes or clinical estimates of a large fetus. In other words, the indications for ultrasound in such studies likely enrich the frequency of birth weight macrosomia, which is the outcome of interest. We are of the view that interpretation of our seemingly good results in the prediction of macrosomia requires some circumspection. Combining readily identifiable clinical risk factors with ultrasound estimates of fetal weight and an increased AFI resulted in a 71% positive predictive value for macrosomia. This value, although higher than described in most other studies, should be viewed in the context that the prediction of macrosomia was also incorrect in 29% of cases. Conversely, in pregnant women without risk factors and without ultrasound estimates suggesting either macrosomia or an increased AFI, 94% can be expected to have neonates weighing less than 4000 g. The bottom line, however, is that simply combining ultrasound findings and clinical features can be quite helpful in the prediction of macrosomia.
Received August 22, 2007, from the Department of Obstetrics and Gynecology, University of Texas Southwestern Medical Center, Dallas, Texas USA. Revision requested September 18, 2007. Revised manuscript accepted for publication September 25, 2007.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Top
Abstract
Introduction
2 prior deliveries), an estimated gestational age of 40 weeks or greater, a male fetus, and Hispanic race. A woman’s prior birth weight history was not available in our data set. 
2, Fisher exact, and Student t tests. Logistic regression analysis was performed to assess the risk for a birth weight of 4000 g or higher, taking into account the clinical risk factors and AFI. Results were considered statistically significant at P 
.05. 