Ultrasonographic Estimation of Fetal Weight

Transcription

1 Article Ultrasonographic Estimation of Fetal Weight Acquiring Accuracy in Residency Mladen Predanic, MD, MSc, Angel Cho, MD, Flores Ingrid, MD, John Pellettieri, MD Objective. Ultrasonographic imaging is considered an objective means for fetal weight estimation. The goals of this study were to determine the accuracy of ultrasonographic estimates of fetal weight performed by residents in training and to ascertain how rapidly the residents gained proficiency in this regard. Methods. A total of 300 ultrasonographic estimates of fetal weight and corresponding birth weight were collected and stratified into 4 groups by the level of residents experience, from level 1 (inexperienced, with <6 months of exposure) to level 4 (advanced experience, with at least 24 months of training). The proportional difference between ultrasonographic estimates of fetal weight and birth weight was calculated for each case and grouped according to the level of training of the examiner. The derived data were compared by analysis of variance, linear regression, and χ 2 test. Results. Significant increases in the accuracy of ultrasonographic estimates of fetal weight were observed with advancing levels of resident experience (P <.0001). Overall, 30.6% of ultrasonographic estimates of fetal weight fell within 5% of birth weight, and 60.6% fell within 10%. Among the least experienced residents (<6 months of training), 49.4% of estimates fell within 10% of birth weight; among those with 6 to 11 months of experience, 53.5% of estimates fell within 10%; among those with 12 to 23 months of experience, 64.1% of estimates fell within 10%; and among the most experienced ( 24 months), 73.6% of estimates fell within 10%. Conclusions. There is a learning curve for ultrasonographic estimates of fetal weight, with a significant decrease in the percent error seen with advancing training among residents, reaching acceptable levels of more than 70% of estimates within 10% of birth weight after 24 months of ultrasonographic experience. Key words: estimated fetal weight; residency; ultrasonography. Abbreviations AC, abdominal circumference; BPD, biparietal diameter; BW, birth weight; EFW, estimates of fetal weight; FL, femur length; uefw, ultrasonographic estimates of fetal weight Received December 4, 2001, from the Department of Obstetrics and Gynecology, The New York Flushing Hospital Medical Center, Flushing. Revision requested January 9, Revised manuscript accepted for publication January 14, We thank E. A. Friedman, MD, DSc, for assisting us in revising the study and guiding us through research principles. Address correspondence and reprint requests to Mladen Predanic, MD, MSc, Department of Obstetrics and Gynecology, The New York Flushing Hospital Medical Center, th Ave, Flushing, NY Estimates of fetal weight (EFW) in late pregnancy are relied on as potentially useful variables for clinical decision making in obstetrics. Clinical EFW based on abdominal examination, including symphysis-fundus height measurements and gestational age, tend to be in a range from accurate to unreliable. 1 3 Ultrasonographic imaging is currently considered sufficiently accurate for objective estimation of fetal weight and clinical applicability. 4 It can serve to determine the weight of the fetus within 10% of actual birth weight (BW) in as many as 75% of estimates and within 5% in as many as 40%. 5 This degree of accuracy is dependent on actual fetal weight; there is a consistent tendency to either overestimate or underestimate EFW at the extremes of the fetal weight range. 3 Errors associated with EFW of small or large fetuses may be harmful if clinical decisions based 2002 by the American Institute of Ultrasound in Medicine J Ultrasound Med 21: , /02/$3.50

2 Acquiring Ultrasonographic Accuracy on such erroneous estimates result in inappropriately premature delivery or lead to surgical delivery in an effort to avert the potential hazards of delivering possible macrosomic infants vaginally. A goal of this study was to determine whether experience improves accuracy in estimating fetal weight among obstetrics and gynecology residents in training. If so, a further objective was to learn how rapidly an acceptable degree of skill can be achieved over the course of the residents training program. Materials and Methods A total of 671 ultrasonographic estimates of fetal weight (uefw) were obtained by residents in the 4-year obstetrics and gynecology program at The New York Flushing Hospital Medical Center during 14 months from September 1998 to December During that period, a total of 2531 women were delivered of infants at our facility. Cases were not specifically assigned to residents at any particular level, nor was any effort made to ensure consistent or random assignments; thus the case selection was a convenience sample. Data were collected in each case for uefw as well as for BW plus maternal and gestational age. In all cases, uefw was obtained on the day of the patient s admission to the labor and delivery unit for a labor-related reason. A total of 300 sequential uefw were included in the study on the basis of the strict inclusion criterion of delivery of the fetus within 48 hr of the uefw at admission. This was applied to ensure against confounding the EFW by continued fetal growth, even though it has been reported that weight estimates are not significantly affected by the period between ultrasonographic examination and delivery if less than 2 weeks. 6 Additionally, we excluded cases in which the fetal head was embedded in the pelvis because of the well-documented reduced accuracy of biparietal diameter (BPD) measurements made on such heads. 7 Any uefw in which all requisite measurements were not recorded were also eliminated from the study. Individual uefw were calculated by the formulas of Shephard et al 8 and Hadlock et al 9 with the use of measurements of BPD and abdominal circumference (AC) 8 or femur length (FL) and AC. 9 Ultrasonographic measurements of BPD, FL, and AC were taken at appropriate fetal anatomic levels, as described elsewhere Each of these results was then compared with the corresponding BW to quantify the error inherent in each measurement. This was calculated as a percentage of the difference divided by BW [(BW uefw)/bw 100]. The mean error of uefw for each stratified group was then determined. In addition, because the raw results obtained by the 2 different formulas were found to be without a statistically significant difference, we used the arithmetic mean to calculate the combined percent error for each EFW. The data thus obtained were stratified by the residents level of experience, defined in 4 levels as level 1 (<6 months of training in ultrasonographic imaging), level 2 (6 11 months), level 3 (12 23 months), and level 4 ( 24 months). A total of 15 residents were involved, none of whom had had any experience in ultrasonographic evaluation of fetal weight before entering the residency program. Although all residents below the level of third and fourth years were instructed and supervised by more experienced personnel, only measurements made independently by them were assessed for purposes of this analysis. During the residency, at the first-year level, ultrasonographic experience would be obtained as the resident observed and worked while being closely supervised by senior personnel in the labor and delivery suite; during the second year, more ultrasonographic experience was gathered during the rotation through the Division of Maternal-Fetal Medicine. The study was started during the second half of the academic year. Therefore, at the beginning of the study, only 12 residents, including 3 residents at each level, were included. However, by the beginning of the first half of the second academic year (the duration of the study was 14 months), 3 additional new residents were included in the study at level 1, and the other residents were advanced into the more experienced ultrasonographic levels accordingly. Imaging studies were carried out using a GE RT3200 Advantage II ultrasonography machine with a 3.5-MHz curved linear probe (GE Medical Systems, Milwaukee, WI). Comparisons between the 2 formulas for obtaining uefw were analyzed by paired t test. The grouped data on mean uefw for each of the experience levels were subjected to statistical analysis by multiple analysis of variance. The trend denoting improvement in accuracy by year of training was examined by time-trend analysis 496 J Ultrasound Med 21: , 2002

3 Predanic et al using linear regression. P.05 was deemed significant. All statistical analysis was performed by GraphPad Prism statistical software, version 3.02 (GraphPad Software, Inc, San Diego, CA). Results Table 1. Relative Accuracy Achieved for uefw Obtained by the Formulas of Shepard et al 8 and Hadlock et al 9 According to Level of Residents Ultrasonographic Experience Mean Error of uefw, % Level of Training Experience Period, mo Shepard et al 8 Hadlock et al 9 1 < ± ± ± ± ± ± ± ± 4.2 Total 8.7 ± ± 5.5 Data are presented as mean ± SD. No statistically significant differences were encountered in pairwise comparison between the data obtained by the 2 formulas (t = 0.52; P =.60), but highly significant differences were shown among the data by experience level (F = 21.3; P <.001). The 300 cases were distributed about equally into 3 of the 4 experience level groups (83, 86, and 72, respectively), but there was a somewhat smaller number (59 measurements) in the group with 12 to 23 months of experience. Nevertheless, the cases in each group were similar with regard to demographic composition. The mean overall maternal age was 26.3 years (range, years). Maternal age did not differ significantly among groups, the age means extending narrowly between 25.9 ± 5.9 and 26.9 ± 6.4 years. The mean gestational age for all cases was 37.3 weeks (range, weeks). In all, 88.3% were at term. The remaining 11.7% of premature births included 3.7% between 24 and 30 weeks and 1.7% between 20 and 24 weeks. The distributions of gestational age were comparable within groupings by ultrasonographic experience. The mean gestational age within groups clustered between 37.3 ± 4.5 and 37.9 ± 3.3 weeks, and the frequency distributions were also similar among them, the rate of premature births at less than 37 weeks ranging from 10.2% to 13.9% (χ 2 = 2.6; df = 9; P =.98). Determination of the mean error of uefw for each of the groupings showed significant improvement with advancing experience (Table 1). Comparing the results according to the 2 different formulas for calculating the EFW, we found that the overall error rates for the residents were 8.7% (range, 0.1% 31.0%) and 8.6% (range, 0.1% 25.0%). There were no demonstrably significant differences between data obtained with the 2 formulas (t = 0.52; P =.60). Therefore, we used the arithmetic mean to calculate the combined percent error for each EFW. When stratified by resident level of ultrasonographic experience, the mean error data among groups were found to be significantly different (F = 21.3; P <.001). Plotting the error data against the actual (ungrouped) duration of experience revealed an apparent diminution in both the spread and degree of error with grater experience (Fig. 1). When these ungrouped data were subjected to linear regression analysis, we were able to verify that accuracy in estimating fetal weight improved significantly over time (F = 20.4; P <.0001). Quantitatively, the degree of improvement could be expressed as a linear regression curve of diminishing error with increasing experience by the equation y = MO, where y is percent error for uefw, and MO is months of ultrasonographic experience. The frequencies of uefw measurements that fell within certain boundaries of error relative to the true BW were next addressed, using cutoff limits of 5% and 10% (Table 2). In all, 60.6% of measurements proved to be within 10% of BW, whereas 30.6% fell within 5%. There was a clear Figure 1. Linear regression for percent error of uewfs and the number of months of residents ultrasonographic experience (P <.0001; slope, ± 0.028; 95% confidence interval, to 0.072; r 2 = 0.06; F = 20.41). J Ultrasound Med 21: ,

4 Acquiring Ultrasonographic Accuracy Table 2. Frequency of uefw Measurements Within Given Accuracy Limits by Resident Experience Level Limits of Measurement Error, % Level of Training Experience Period, mo <5% <10% >10% 1 < Total Data are expressed as percent measurements within the designated limits. trend toward increasing accuracy with the use of both of these limits. Data were within 10% in just less than half of the cases for inexperienced residents but rose to nearly three fourths for the most experienced, whereas accuracy within 5% of BW rose steadily from less than one fifth to more than twice that with benefit of experience. Discussion Accurate estimation of fetal weight remains an elusive goal for obstetricians. Recent published studies have embraced the concept that ultrasonography is valuable for detecting fetal macrosomia, with the aim of reducing the incidence of shoulder dystocia and the complications associated with it Although both clinical and ultrasonographic estimation of fetal weight greater than 4000 g BW seem acceptably accurate, 2 our ability to diagnose fetal macrosomia using ultrasonography is quite limited. 16 Clinical evaluation of EFW in term pregnancies may actually be more accurate than ultrasonographic imaging. 3 Nonetheless, the best results for ultrasonographic determination of fetal weight are encountered if the fetus weighs between 2500 and 3500 g, and the worst are obtained with fetuses weighing more than 4000 g. 17 As to possible explanations for this poor correlation between uefw and BW at the extremes of BW, 1 it may be conjectured that the equations used are insufficiently accurate, because they were derived from populations of fetuses within a narrow range of BW. If so, they would not necessarily be appropriate for use outside that range. Alternatively, they may not be applicable to all ethnic groups. Similarly, fetuses of diabetic gravidas or those with intrauterine growth restriction for whatever reason may require different formulas for estimating their weight. Several studies tried to evaluate the influence of maternal obesity and oligohydramnios as possible causes of poor ultrasonographic accuracy at term. Some found that maternal weight and low amniotic fluid index did not affect ultrasonographic accuracy, 6,18,19 whereas others showed that oligohydramnios tended to increase the mean error. 20,21 This latter finding suggested that intrapartum EFW should be obtained before artificial rupture of the membranes. 22 Variable ultrasonographic accuracy for assessing fetal weight can be explained by at least 2 critical sources of error in the calculation of uefw. 23 First, the measurement is based on formulas that use 2 or more variables (BPD and either FL or AC) and assume a uniform relationship between two-dimensional area derived from those diameters and volume, which is then translated into weight. That relationship may not actually exist; it is in fact unlikely to exist, given that few if any of the objects being assessed are perfectly globular structures. Second, estimation of fetal weight assumes that there is a constant relationship between volume and mass (density) among all fetuses. Structural weight for a given volume, however, can vary considerably according to its water and fat content. The smallest errors are achieved when the fetal volume is measured by a technique of water displacement or by using measurements derived from serial ultrasonographic planes (yielding 7.6% and 6.5% errors, respectively). 24,25 Nonetheless, because BPD, FL, and AC measurements are so much easier to master and have become the clinical standard for making such measurements, we accepted their limitations and used them in this study. We tried to evaluate the role of residents experience and the skill levels they achieved in estimating fetal weight by the ultrasonographic imaging technique over the course of their education in our residency training program. Our results showed that the mean error was greatest during the first several months of the residents introduction to the method and that it gradually improved during their training. There were no differences between the results derived from the 2 formulas used, making either equally reliable for estimating fetal weight. Although the mean error ranged from 10.8% to 6.8% (levels 1 4), the mean proportions of measurements with errors within 5% and 10% of actual BW for all levels of experience were 30.6% and 60.6%, respectively. This is less than previously published studies, in 498 J Ultrasound Med 21: , 2002

5 Predanic et al which 40% and 75% of measurements fell within 5% and 10% margins of error. 5 Other studies reported even better results, with 50% of uefw measurements falling within 5% 19 and 79% falling within 10%. 6 By the time our residents had sufficient experience of more than 12 months to achieve substantial skill, the proportion of estimates within 10% of BW rose to 64.1% and was 73.6% after 24 months of training. A limit of ±10% for ultrasonographic EFW, which is considered optimal and acceptable for prospective treatment of patients in labor, 3 was reached by residents over time, showing that training was effective. Estimates of fetal weight obtained by residents toward the end of their residency program proved both reliable and fairly predictive of actual BW and were comparable with accuracy rates previously reported. 3,6,19 In conclusion, there is a learning curve for uefw. In this study, it did not seem to be a single point at which there was a statistically significant improvement, but the improvement was gradual over the resident training period. However, if an acceptable level of accuracy is determined to be more than 70% of estimates within 10% of BW, than a significant improvement is accomplished after 24 months exposure to ultrasonographic training. References 1. Mongelli M. Fetal weight estimation by symphysisfundus height and gestational age. Gynecol Obstet Invest 1997; 43: Raman S, Urquhart R, Yusof M. Clinical versus ultrasound estimation of fetal weight. Aust N Z J Obstet Gyneacol 1992; 32: Chauhan SP, Hendrix NW, Magann EF, Morrison JC, Kenney SP, Devoe LD. Limitations of clinical and sonographic estimates of birth weight: experience with 1034 parturients. Obstet Gynecol 1998; 91: American Collage of Obstetricians and Gynecologists. Fetal Macrosomia. Washington, DC: American College of Obstetricians and Gynecologists; Educational Bulletin Watson WJ, Soisson AP, Haarlass FE. Estimated weight of the term fetus: accuracy of ultrasound vs. clinical examination. J Reprod Med 1988; 33: Kaaij MW, Struijk PC, Lotgering FK. Accuracy of sonographic estimates of fetal weight in very small infants. Ultrasound Obstet Gynecol 1999; 13: Shamley KT, Landon MB. Accuracy and modifying factors for ultrasonographic determination of fetal weight at term. Obstet Gynecol 1994; 84: Shepard MJ, Richards VA, Berkowitz RL, Warsof SL, Hobbins JC. An evaluation of 2 equations for predicting fetal weight by ultrasound. Am J Obstet Gynecol 1982; 142: Hadlock FP, Harrist RB, Carpenter RJ, Deter RL, Park SK. Sonographic estimation of fetal weight: the value of femur length in addition to head and abdomen measurements. Radiology 1984; 150: Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal biparietal diameter: a critical re-evaluation of the relation to menstrual age by means of real-time ultrasound. J Ultrasound Med 1982; 1: Hadlock FP, Harrist RB, Deter RL, Park SK. Fetal femur length as a predictor of menstrual age: sonographically measured. AJR Am J Roentgenol 1982; 138: Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal abdominal circumference as a predictor of menstrual age. AJR Am J Roentgenol 1982; 139: Levine AB, Lockwood CL, Brown B, Lapinski R, Berkowitz RL. Sonographic diagnosis of large for gestational age fetus: does it make a difference? Obstet Gynecol 1992; 79: Deter R, Hadlock F. Use of ultrasound in the detection of macrosomia: a review. J Clin Ultrasound 1985; 13: Delpapa EH, Muller-Heubach E. Pregnancy outcome following ultrasound diagnosis of macrosomia. Obstet Gynecol 1991; 78: Gonen R, Spiegel D, Abend M. Is macrosomia predictable, and are shoulder dystocia and birth trauma preventable? Obstet Gynecol 1996; 88: Acker DB, Sach BP, Ransil BJ, Friedman EA. Ultrasonography for fetal weight estimation: the Birnholz equation. Ultrason Imaging 1987; 9: Blann DW, Prien SD. Estimation of fetal weight before and after amniotomy in the laboring gravid J Ultrasound Med 21: ,

6 Acquiring Ultrasonographic Accuracy woman. Am J Obstet Gynecol 2000; 182: Field NT, Piper JM, Langer O. The effect of maternal obesity on the accuracy of fetal weight estimation. Obstet Gynecol 1995; 86: Barnhard Y, Bar-Hava I, Divon MY. Accuracy of intrapartum estimates of fetal weight: effect of oligohydramnios. J Reprod Med 1996; 41: Edwards A, Goff J, Baker L. Accuracy and modifying factors of the sonographic estimation of fetal weight in a high-risk population. Aust N Z J Obstet Gynaecol 2001; 41: Dar P, Weiner I, Sofrin O, Sachs GS, Bukovsky I, Arieli S. Clinical and sonographic fetal weight estimates in active labor with ruptured membranes. J Reprod Med 200; 45: Manning FA. General principles and applications of ultrasonography. In: Creasy RK, Resnik R. Maternal- Fetal Medicine. 4th ed. Philadelphia, PA: WB Saunders Co; 1999: Thompson TR, Manning FA. Estimation of volume and weight of the perinate: relationship to morphometric measurement by ultrasonography. J Ultrasound Med 1983; 2: McCallum WD, Brinkley TF. Estimation of fetal weight from the ultrasound measurements. Am J Obstet Gynecol 1979; 123: J Ultrasound Med 21: , 2002