| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the University of Rochester, Rochester, New York.
Address reprint requests to: Gail Best, MD, University of Louisville, Department of Obstetrics and Gynecology, Ambulatory Care Building, Louisville, KY 40292.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: Patients with singleton pregnancies who had diabetes and who underwent sonograms between 34.0 weeks and 36.9 weeks formed the study group. The control group was comprised of nondiabetic women who had sonograms in the same gestational age range. The absolute birth weight errors and absolute percent errors in the study group and the controls were compared with t tests.
RESULTS: A total of 133 diabetic women and 1690 controls were included in the study. The mean (± standard deviation) absolute error of the predicted birth weight was 265 ± 210 g in the diabetic women and 261 ± 204 g in the control group (P = .87). The mean (± standard deviation) absolute percent error was 7.4% ± 6.3% for diabetic women compared with 8.3% ± 6.6% for the controls (P = .14).
CONCLUSION: Prediction of birth weight using the gestation-adjusted projection method is as accurate in diabetic women as in controls.
Macrosomia, defined as a birth weight of greater than 4000 g, occurs in 25–42% of diabetic pregnancies.1 This is clinically significant, as macrosomia is associated with significant neonatal and maternal morbidities, particularly in diabetic pregnancies. Macrosomia can result in trauma to the neonate and mother, asphyxia, and an increased risk of operative delivery and postpartum hemorrhage. Currently available methods to estimate birth weight have limitations and are less accurate in macrosomia.2,3 If birth weight could be accurately estimated, this information would be extremely helpful in determining the appropriate route of delivery, particularly in diabetic women.
Most studies of birth weight estimates have been performed within 1 week of delivery. This approach minimizes the effects of subsequent fetal growth on ultrasound examinations. However, at this stage, there has generally been descent of the presenting part into the maternal pelvis. The measurement of biparietal diameter and head circumference when descent has occurred is often difficult if not impossible. Additionally, at this time in gestation, there is a relative decrease in the amount of amniotic fluid, which may also affect the accuracy of measurements.4–6
To overcome these limitations, several authors have described techniques for the forward extrapolation of sonographic estimates of fetal weight to the time of delivery.7–9 The gestation-adjusted projection method is one such technique. It is based on the hypothesis that the ratio of estimated fetal weight to the median fetal weight for gestational age remains constant in the third trimester.9 We previously examined estimates of birth weight by ultrasounds performed at 34–36.9 weeks and compared them with those performed in the same patients at 37 weeks’ gestation or later. Using gestation-adjusted projection, the mean absolute error for predicting birth weight was lower when the earlier ultrasound examination was used to predict birth weight (6.2% versus 7.4%).10 The study sample in this previous report included diabetic patients, but there were insufficient numbers to evaluate this technique in diabetic patients separately from the entire study group.
The purpose of the present investigation was to compare diabetic and nondiabetic women near term with respect to the accuracy of predicted birth weight by the gestation-adjusted projection method. Because gestation-adjusted projection applied at 34–37 weeks’ gestation showed a small but statistically significant improvement in birth weight prediction over the same method applied later in pregnancy in a general population, we hypothesized that this method at this gestational age may provide a more accurate way to estimate birth weight in a population at high risk for delivery-related complications (ie, diabetic women).
|
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
All ultrasound examinations were performed by an experienced sonographer using standard techniques.11 The studies were completed on one of the following ultrasound units: ATL 5000, ATL 3500, ATL 3000, GE 400, AI 5200, HDI Ultramark 9, or an ATL Ultramark 4 (ATL, Philips Medical Systems, Best, the Netherlands; GE, GE Medical Systems, Fairfield, CT; HDI, Philips Medical Systems, Best, the Netherlands; AI, Dornier Medical Systems, Phoenix, AZ). Fetal weight was calculated using Hadlock’s formula.12,13
Data were collected on fetal biometry, estimated fetal weight, gestational age at the time of ultrasound, gestational age at delivery, birth weight, and the latency period from the time of the ultrasound to the time of delivery.
Birth weight for each patient was predicted from the results of the ultrasound using the gestation-adjusted projection method and Brenner’s median fetal weights for gestational age.14 In brief, the ratio between the estimated fetal weight and the median fetal weight for the gestational age at that sonogram was calculated. This ratio was then multiplied by the median birth weight at the gestational age of delivery to give the predicted birth weight (Figure 1
). A graphic representation of this method is presented in Figure 2. The bold line represents the median fetal weight for gestational age. The percentiles above and below the median are also represented. Similar to childhood growth curves, an estimated fetal weight can be plotted on this graph and the appropriate percentile curve followed to the gestational age at delivery to estimate the predicted birth weight.
|
|
2 test was used to compare these proportions between the two groups. A P value of .05 was considered statistically significant.
A power calculation was performed to determine the number of patients needed to show a 100-g difference in absolute error of birth weight prediction. Using an 
of 0.05 and a ß of 0.9, an 8% incidence of diabetes in our population and a variance of 151.9 g from our prior study in a general population, 583 subjects (47 diabetic women and 536 controls) would be required.
|
RESULTS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Mean birth weight, mean latency, mean absolute error, and mean absolute percent error are presented in Table 1. There was no difference in the mean absolute error of predicted birth weight between the study and control groups, whereas the mean absolute percent error was less in the diabetic population. A normal distribution was found to be present after logarithmic transformation of the absolute percent errors was performed. There was no difference between the two groups in the log-transformed mean absolute percent error. Percentage of correct birth weight predictions is presented in Table 2, again suggesting a somewhat better accuracy for diabetic women than controls.
|
|
. The gestation-adjusted projection method appeared to perform equally well for diabetic women as for normal controls.
|
|
DISCUSSION |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
There are three limitations to our study. The first is that the subjects included were not recruited specifically for this study. The sonographic data were obtained from our database of patients referred for a variety of indications. The nondiabetic patients referred for sonograms between 34.0 and 36.9 weeks’ gestation may not be representative of a normal population because they had some indication for an ultrasound examination.
The second possible limitation is that patients were not matched for variables that may affect the accuracy of estimated fetal weight, such as maternal body mass, amount of amniotic fluid, and fetal presentation. Visualization can be quite limited, depending on maternal body habitus, making determination of fetal dimensions difficult. The effect of amniotic fluid volume on the accuracy of fetal weights is unclear. In the presence of low amniotic fluid volume, sonographically determined estimated fetal weight has been found to be significantly less accurate.4,5 However, in two other studies, the amount of amniotic fluid was not shown to significantly impact on accuracy of the estimated fetal weight.2,15 The accuracy of fetal weight estimation can also be affected by an abnormal cephalic index.5 Fetal presentation may alter the shape of the fetal head and influence the accuracy of measurements.
A third limitation to the study is that maternal complications were not always present or completely characterized in the ultrasound database. As such, a separate evaluation of the gestation-adjusted projection method in pregestational and gestational diabetic patients could not be performed. Additionally, maternal complications that could affect fetal growth were not recorded for all patients.
Fetuses thought to be macrosomic were evaluated independently. Sonographic screening has been found to be poor at accurately predicting macrosomia in nondiabetic women at term.3,16 Pedersen and Molsted-Pedersen17 evaluated the accuracy of fetal weight estimation in diabetic women using abdominal circumference alone. They found that the positive and negative predictive values of predicting macrosomia using this method were 80% and 93%, respectively.17 Other methods at detecting macrosomia have been evaluated. Cheek-to-cheek diameter has been reported to improve the accuracy of estimated fetal weight in macrosomic fetuses.18
In our study, the prediction of macrosomia in the diabetic group was actually more accurate than in the women who were not diabetic. The mean absolute percent error was 6.8% in diabetic subjects and 10.1% in nondiabetic subjects. There was a higher positive predictive value in the diabetic group than in the nondiabetic group. These findings may reflect the higher incidence of macrosomia in the diabetic group as well as the higher mean birth weight in this group, rather than a true improved accuracy in the prediction of birth weight.
The apparent improved accuracy in estimated fetal weight prediction in the entire diabetic study group may also be attributed to the higher birth weight in diabetic patients in general. The mean absolute error was similar in the study and control populations. A similar absolute error actually represents a larger percentage of lower birth weight. As such, the mean absolute error of 265 g with a mean birth weight of 3660 g yielded a 7.42% absolute error in the diabetic subjects. By contrast, the mean birth weight in the control group was 3242 g; the absolute error of 262 g in this nondiabetic group produced an 8.3% error.
It has previously been shown that a latency period of greater than 35 days did not affect the accuracy of birth weight prediction in a study that included latency periods up to 49 days.10 Our control population had a longer mean latency period and a wider range, up to 56 days. The difference in latency between the two groups occurred because diabetic patients generally are not delivered past their due date in our institution. It is possible that the accuracy of birth weight prediction may be diminished when the latency interval extends beyond 49 days. The improvement in prediction errors in the diabetic group may also be accounted for by this difference.
The accurate estimation of fetal weight is a challenging, yet clinically important, task. Given the increased risk of complications with macrosomia, particularly in diabetic women, improving methods for determining fetal weight would provide clinicians with a better guide for management of these patients. Our study shows that in diabetic women, the accuracy of fetal weight estimation near term is at least equal to that in women who are not diabetic. Use of the gestation-adjusted projection method allows improved accuracy of prediction over those previously reported. The incorporation of the graphic representation of the gestation-adjusted projection method (Figure 2
) into clinical care may assist in birth weight prediction.
|
Footnotes |
|---|
Received August 6, 2001. Received in revised form November 28, 2001. Accepted December 11, 2001.
|
REFERENCES |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Benacerraf BR, Gelman R, Frigoletto FD. Sonographically estimated fetal weights: Accuracy and limitation. Am J Obstet Gynecol 1988;159:1118–21.[Medline]
3. Ocer F, Kaleli S, Budak E, Oral E. Fetal weight estimation and prediction of fetal macrosomia in non-diabetic pregnant women. Eur J Obstet Gynecol 1999;83:47–52.[Medline]
4. Barnard Y, Itai BH, Divon MY. Accuracy of intrapartum estimates of fetal weight: Effects of oligohydramnios. J Reprod Med 1996;41:907–10.[Medline]
5. Ott WJ, Doyle S, Flamm S. Accurate ultrasonic estimation of fetal weight. Am J Perinatol 1986;3:193–7.[Medline]
6. Rossavik IK, Joslin GL. Macrosomatia and ultrasonography: What is the problem? Southern Med J 1993;86: 1129–32.[Medline]
7. Spinnato JA, Allen RD, Mendenhall HW. Birth weight prediction from remote ultrasound examination. Obstet Gynecol 1988;71:893–8.[Medline]
8. Spinnato JA, Allen RD, Mendenhall HW. Birth weight prediction from remote ultrasonographic examination. Am J Obstet Gynecol 1989;161:742–7.[Medline]
9. Mongelli M, Gardosi J. Gestation-adjusted projection of estimated fetal weight. Acta Obstet Gynecol Scand 1996; 75:28–31.[Medline]
10. Pressman EK, Bienstock JL, Blakemore KJ, Martin SA, Callan NA. Prediction of birth weight by ultrasound in the third trimester. Obstet Gynecol 2000;95:502–6.
11. DuBose T. Assessment of fetal age and size: Techniques and criteria. In: Berman MC, Cohen HL, eds. Diagnostic medical sonography, a guide to clinical practice, obstetrics and gynecology. 2nd ed. Philadelphia: Lippincott; 1997: 359–98.
12. Hadlock FP, Harrist RB, Sharman RS, Deter RL, Park SK. Estimation of fetal weight with the use of head, body and femur measurements — A prospective study. Am J Obstet Gynecol 1985;151:333–7.[Medline]
13. Hadlock FP, Harrist RB, Martinez-Poyer J. In utero analysis of fetal growth: A sonographic weight standard. Radiology 1991;181:129–33.
14. Brenner WE, Edelman DA, Hendricks CH. A standard of fetal growth for the United States of America. Am J Obstet Gynecol 1976;126:555–64.[Medline]
15. Patterson RM. Estimation of fetal weight during labor. Obstet Gynecol 1985;65:330–2.
16. Pollack RN, Hauer-Pollack G, Divon MY. Macrosomia in postdates pregnancies: The accuracy of routine ultrasonographic screening. Am J Obstet Gynecol 1992;167:7–11.[Medline]
17. Pedersen JF, Molsted-Pedersen L. Sonographic estimation of fetal weight in diabetic pregnancy. Br J Obstet Gynaecol 1992;99:475–8.[Medline]
18. Abramowicz JS, Robischon K, Cox C. Incorporating sonographic cheek-to-cheek diameter, biparietal diameter and abdominal circumference improves weight estimation in the macrosomic fetus. Ultrasound Obstet Gynecol 1997; 9:409–13.[Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
