International Journal of Gynecology & Obstetrics 73 Ž2001. 15᎐20 Article Assessment of fetal weight based on ultrasonic femur length after the second trimester 夽 Mehrdad Honarvar U , Mahboobeh Allahyari, Sedigheh Dehbashi Department of Obstetrics and Gynecology, Shiraz Uni¨ ersity of Medical Sciences, Shiraz, Iran Received 8 August 2000; received in revised form 23 October 2000; accepted 25 October 2000 Abstract Objecti¨ e: The ultrasonic measurement of the fetal femur length is a sensitive and precise variable for estimation of fetal growth and development. The objective of this study is to predict fetal weight in fetuses of more than 24 weeks gestation by ultrasound measurement of the femur length. Method: In this study, pregnant mothers were identified by the criteria of normalities, such as: well-known LMP, regular menstrual cycles, no use of OCP for the last 3 months, no use of alcohol or cigarettes, no drug abuse, no history of diabetes or chronic HTN. Multiple gestations, congenital anomalies and still-births were excluded. Birth-weight measurements Žadjusted for maternal age, baby’s sex, parity and week of gestation. were taken immediately after birth. Result: The relation between fetal weight and fetal femur length has been determined by cross-sectional analysis of 900 normal fetuses ŽG 25 weeks gestation. using real time ultrasonography. Mathematical modeling of the data has demonstrated that the femur growth curve is non-linear beyond 24 weeks gestation. With the aid of a scientific calculator the data were analyzed and a simple second-grade equation has been derived: EFW Žkg. s 0.042FL2 Žcm. q 0.32FLy 1.36, S.D.; "235 g ŽHonarvar’s Formula 2.. With the use of this data, the error in estimation of EFW given FL is "235 g. Conclusion: This simple, new and accurate equation appears to be clinically reliable and easy to use and suggests that previous 夽 This paper was presented in part at the First Asian Pacific International Congress of Anatomists Ž1st APICA., Seoul, South Korea, 22᎐24 August 1996 and the 15th World Conference of Family Doctors ŽWONCA98., Dublin, Ireland, 14᎐18 June 1998 and the First Emirates International Congress of Perinatology Ž1st EICOP., Abudhabi, United Arab Emirates, 9᎐12 November 1998. U Corresponding author. E-mail address: alziba@yahoo.com ŽM. Honarvar.. 0020-7292r01r$20.00 䊚 2001 International Federation of Gynecology and Obstetrics. All rights reserved. PII: S 0 0 2 0 - 7 2 9 2 Ž 0 0 . 0 0 3 6 8 - 4 16 M. Honar¨ ar et al. r International Journal of Gynecology & Obstetrics 73 (2001) 15᎐20 normal ultrasonic fetal femur length curves for another population may underestimate or overestimate normal fetal weight for the Iranian population. Thus, our formula is an excellent means to estimate true fetal weight. 䊚 2001 International Federation of Gynecology and Obstetrics. All rights reserved. Keywords: EFW, estimated fetal weight; GA, gestational age; FL, femur length 1. Introduction Determination of fetal weight is important in all pregnancies. When the delivery of an infant is anticipated, accurate antenatal assessment of the fetal weight can be a useful adjuvant for establishing a plan of management that will minimize the perinatal morbidity and mortality rate w1x. A major problem in the decision-making process seems to be the inability to estimate fetal weight accurately prior to delivery w2x. Recently, a number of equations for estimating fetal weights, based on ultrasound measurements obtained within 48᎐72 h prior to delivery, have been devised w3᎐5x and tested. Besides it is possible to identify abnormal fetal growth with the measurement of normal growth curves for femur length during pregnancy. It is noticed that the growth rate of femur length illustrates a decreasing growth rate as gestational age increases. This growth rate can be used in following the progressive growth of the fetus w6x and infants who deviate from physiologic norms of FL for GA have increased perinatal mortality and morbidity. Thus, symmetric growth retardation can be identified in all instances by shortened ultrasonic femur lengths. On the other hand, the fairly common practice of using mean values determined for one population as the basis for judging the individual fetus drawn from another population compounds the problem of lack of sensitivity in two ways. It ignores the fact that all normal fetuses within a population do not conform to mean values and it does not take into account interpopulation variation in fetal anthropometric characteristics. For example, Raman et al. w7x have shown a significant difference in limb length growth of Indians compared with non-Indians. For this reason, the present report is a detailed account of our experi- ence with estimating EFW in fetuses of more than 24 weeks gestation in an Iranian population. In the Iranian population this is one major ethnic group. Our aim was to make a prospective study of growth of the femur in the Iranian ethnic group and observe any differences in growth vs. other populations. Thus, we intend to device a growth chart of femur lengths applicable for clinical use in our local setting. 2. Methods Gestational age assessment in our study was performed by ultrasonography by using crown᎐rump length. CRL was measured in the first trimester Žbefore the 12th week of gestation. to date of pregnancy. Therefore, in each case gestational age was documented by the last menstrual period and correlated with ultrasonic CRL measurements to within "5 days. Pregnant mothers participated in this study demonstrated criteria of normality such as: 1. regular menstrual cycles, at least for three menses; 2. no use of OCP, for at least 3 months previous; 3. live-born neonates; 4. no anomaly known perinatally or postnatally; 5. no use of alcohol or cigarettes; 6. no drug abuse; 7. no family H.X of dwarfism; and 8. no diabetes or chronic HTN in mothers. The patients who have the following criteria were excluded: 1. multiple gestations Ži.e. twins.; 2. gestational duration ) 40 week; M. Honar¨ ar et al. r International Journal of Gynecology & Obstetrics 73 (2001) 15᎐20 Table 1 Obstetrics data for normal patients Ž N s 900. Demographic data Parity Para 0 Para 1 Para 2 or more Number % Male neonates 374 256 270 41.5 196 28.5 121 30 126 Female neonates 178 135 144 Economical class Žmonthly income. Low Ž- $100. 72 8 Middle Ž$100᎐$300. 702 78 High Ž) $300. 126 14 Mean patient age 23.78 years 3. maternal chronic disease; and 4. non-Iranian ethnic groups. Following delivery, neonatal actual birth weights ŽABWs. were immediately obtained and recorded. The patient population is described in Table 1. This study consists of 900 pregnant women and 1062 femur length measurements were obtained, 162 of these were studied serially and had at least two measurements made throughout their preg- 17 nancies. In these patients the femur length measured just before birth was the one used to correlate with birth weight. The FL is defined from the proximal to the distal metaphysis, obtained within 48᎐72 h prior to deli¨ ery. For measuring FL, the present study utilized the technique originally described by O’Brien et al. w6x. In this connection, the long axis of the fetus is identified first and the transducer is then turned 90⬚ to produce a cross-sectional image of the fetal trunk. After that several femur length measurements are made, within a 2-mm range, and the average measurement is considered optimal. All examinations were performed by one of us using a linear-array real-time system with a 3.5-MHz focused transducer ŽHitachi EUB 40.. The distribution of FL for each week of gestational age is presented in Table 2. Our research was conducted as a longitudinal prospective study of a ‘normal’ population referred from the prenatal centers at the urban and rural areas to the large university hospital in Shiraz. Therefore, our measurements taken in this referral center are indicative of a diverse population and represent the entire Iranian population, not just the population around Shiraz. Table 2 Distribution of femur length for each week of gestational age Week y2 S.D. Žmm. y1 S.D. Žmm. Mean FL Žmm. q1 S.D. Žmm. q2 S.D. Žmm. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 39.48 40.3 41.7 43 51 48.75 50.52 53.77 53.84 57.46 56.46 60.4 58.93 65.22 65.64 68.42 41.54 42.9 44.75 46.2 51.75 51.72 53.66 55.38 56.72 60.13 60.08 63.1 62.86 66.56 67.17 70.26 43.6 45.5 47.8 49.4 52.5 54.7 56.8 58.3 59.6 62.8 63.7 65.8 66.8 67.9 68.7 72.1 45.66 48.1 50.85 52.6 53.25 57.67 59.94 61.21 62.48 65.47 67.32 68.5 70.73 69.24 70.23 73.94 47.72 50.7 53.9 55.8 54 60.65 63.08 64.13 65.36 68.14 70.94 71.2 74.67 70.58 71.76 75.78 M. Honar¨ ar et al. r International Journal of Gynecology & Obstetrics 73 (2001) 15᎐20 18 Fig. 1. EFW against FL. 3. Results In this research the equation for EFW as a function of FL was generated from the data by means of the regression analysis method ŽFig. 1.. From 25 to 40 weeks gestation the correlation of fetal femur length growth with fetal weight seems best to be approximated by a non-linear relationship w7x. So, at this period the linear quadratic function is the optimal model for predicting fetal weight from femur length Ž r s 0.98, S.D.s "235 g.. With the aid of a scientific calculator the data were analyzed and the regression equation for Table 3 Distribution of estimated fetal weight for mean femur lengths at each week of gestational age EFW Žkg. s 0.042FL2 Žcm. q 0.32FLy1.36, S.D.; "235 g ŽHonarvar’s Formula 2. a Week y2 S.D. y1 S.D. Mean q1 S.D. q2 S.D. FL Žcm. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 355 496 659 770 1000 1170 1337 1455 1567 1829 1908 2085 2177 2272 2348 2657 590 731 894 1005 1235 1405 1572 1690 1802 2064 2143 2320 2412 2507 2583 2892 825 996 1129 1240 1470 1640 1807 1925 2037 2299 2378 2555 2647 2742 2818 3127 1060 1201 1364 1475 1705 1875 2042 2160 2272 2534 2613 2790 2882 2977 3053 3362 1295 1436 1599 1710 1940 2110 2277 2395 2507 2769 2848 3025 3117 3212 3288 3597 4.36 4.55 4.78 4.94 5.25 5.47 5.68 5.83 5.96 6.28 6.37 6.58 6.68 6.79 6.87 7.21 a " 1 S.D.s "235; "2 S.D.s "470. M. Honar¨ ar et al. r International Journal of Gynecology & Obstetrics 73 (2001) 15᎐20 Table 4 A comparison of the mean ABWs to our predicted weights Week Mean ABWs EFW da 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Mean 762.5 1031.28 1084.54 1326.92 1344.66 1562.22 1810.00 1985.29 2149.66 2279.68 2410.25 2630.10 2733.13 2870.78 2940.64 3250.52 825 966 1129 1240 1470 1640 1807 1925 2037 2299 2378 2555 2647 2742 2818 3127 y62.5 65.28 44.46 86.92 y125.34 y77.78 y3 60.29 112.66 19.32 32.25 75.1 86.13 128.78 122.64 123.52 22.63 a ds difference ABWs and EFW ŽABWsy EFW.. Kolmogoro¨ ᎐Smirno¨ Z two-sample test: d was not statistically significant ŽK᎐S Z s 0.436, P; 0.991.. our data is: EFW Ž kg. s 0.042FL2 Ž cm. q 0.32FLy 1.36, S.D.; "235 g Ž Honarvar’s Formula 2 . where, EFW is in weeks ŽG 25 weeks gestation. and FL is in centimeters. The S.D. tends not to increase as femur length increases. Predicted EFW values of our study for specific FL measurements are presented in Table 3. After that Kolmogorov᎐Smirnov Z two-sample test for goodness of fit between EFW and mean ABWs was performed and differences were not statistically significant ŽK᎐S Z s 0.436, P; 0.991.. The comparison of mean ABWs and that of EFW derived from Honarvar’s Formula is shown in Table 4. There was no difference in either the distribution of parity or in age between patients who delivered male infants and those delivering female infants. It was found that gender seems not to affect limb lengths w4x. There were 374 patients who were nulliparous, 256 patients primiparous and 270 patients para 2 or greater. The mean age for all patients was 23.78 years ŽTable 1.. 19 4. Discussion When there is a need for deciding on the viability of the fetus or its chances of survival, evaluation of fetal weight seems to be of great importance. The problem of the large fetus may arise in conjunction with abnormalities of labor and with fetuses in breech presentations. Many centers now believe that the large fetus or the term fetus in breech presentation may benefit from operative delivery w4x. The mean delivery weights in our study were compared with that of two normal ultrasonic fetal weight tables. So that, an analysis of the mean delivery weights in our study confirms the findings of Hadlock w8x and Ott w9x in part. The estimates of fetal weight by means of predicted normal ultrasonic weight performed by Hadlock are more likely to be lower than actual neonatal weight. Conversely, the estimates of fetal weight performed by Ott were more likely to be an overestimate of the actual birth weight w10x. After 29 weeks of gestation, the increase in the confidence interval with increasing fetal weight is seen, but before 29 weeks of gestation the confidence interval is variable. The absolute mean error involved in the calculation of fetal weight by ultrasonic methods ranges from 5 to 8.4% w11x. There are major controversies in the reported gestational age variability Ž"2 S.D.. during the period 23᎐40 weeks. For example, Hadlock w12x reported that variability during this period is Ž"3᎐3.5 weeks., but Yeh et al. w13x reported a 95% confidence interval of "5 days. Our study, based on analysis of over 900 fetuses, indicates that the variability during this period is "5 days w14x. Besides, recently some others have shown that the correlation coefficient of gestational age ŽGA. vs. FL is statistically greater than that of the GA vs. BPD w7,15x. Finally, this article suggests that previous normal ultrasonic fetal femur length curve of one population is unsuitable and inappropriate for another population. The estimated quadratic growth curves of femoral length amongst the three ethnic groups clearly showed that the rate of growth of the Indian fetuses was faster than that 20 M. Honar¨ ar et al. r International Journal of Gynecology & Obstetrics 73 (2001) 15᎐20 of the Malays and Chinese Žnon-Indian fetuses .. Their femur lengths grew faster by 0.15 mmrweek, which is highly significant Ž P- 0.001.. The growth curve was, however, similar amongst the three ethnic groups w7x. Therefore, it is recommended that one should not use mean values determined for one population as the basis for judging the individual fetus drawn from another population, because of the significant interpopulation variation observed in fetal femur lengths and this should be taken into account when growth charts are being used and fetal weight formulae are being calculated using limb lengths. In conclusion, in each population, fetal femur length must be studied independently for making a better operational and functional decision in the field of obstetrics and gynecology. w5x w6x w7x w8x w9x w10x w11x References w12x w1x Thurnau GR, Tamura RK, Sabbagha R, Dep III OR, Dyer A, Larkin R et al. A simple estimated fetal weight equation based on real-time ultrasound measurements of fetuses less than thirty-four week’s gestation. Am J Obset Gynecol 1983;145:557. w2x Key TC, Dattel BJ, Resink R. The ultrasonographic estimation of fetal weight in the very low birth-weight infant. Am J Obstet Gynecol 1983;145:574. w3x Warsof SL, Gohari P, Berkowitz RL, Hobbins JC. The estimation of fetal weight by computer assisted analysis. Am J Obstet Gynecol 1977;128:881. w4x McCallun WD, Brinkley JE. Estimation of fetal weight w13x w14x w15x from ultrasonic measurements. Am J Obstet Gynecol 1979;133:195. Shepard MJ, Richards VA, Berkowitz RL, Warsof SL, Hobbins JC. An evaluation of two equations for predicting fetal weight by ultrasound. Am J Obstet Gynecol 1982;142:47. O’Brien GD, Queenan GT. Growth of the ultrasonic fetal femur length during normal pregnancy. Am J Obstet Gynecol 1981;14Ž7.:833. Raman S, Teoh T, Nagaraj S. Growth patterns of the humoral and femur length in a multiethnic population. Int J Gynecol Obstet 1996;54:143. Hadlock FP, Deter RL, Harrist RB. Sonographic detection of fetal intrauterine growth retardation. Appl Radiol 1983;12:28. Ott WJ. The diagnosis of altered fetal growth. Obstet Gynecol Clin North Am 1988;15Ž2.:237. Honarvar M, Allahyari M, Dehbashi S. A simple estimated fetal weight equation for fetuses between 24 and 34 weeks of gestation. Int J Gynecol Obstet 1999;67:67. Ott WJ, Doyle S. Normal ultrasonic fetal weight curve. Obstet Gynecol 1982;59:603. Hadlock FP, Harrist RB, Deter RL et al. Fetal femur length as a predictor of menstrual age: sonographically measured. Am J Roentgenol 1982;138:875. Yeh M, Bracero L et al. Ultrasonic measurement of the femur length as an index of fetal gestational age. Am J Obstet Gynecol 1982;144:519. Honarvar M, Allahyari M, Dehbashi S. Assessment of gestational age based on ultrasonic femur length after the first trimester; a simple mathematical correlation between gestational age and femur length. Int J Gynecol Obstet 2000;70Ž3.:335. Owen P, Donnet ML et al. Standards for ultrasound fetal growth velocity. Brit J Obstet Gynaecol 1996;103:60.