- Email: [email protected]
Identification of the small for gestational age fetus with the use of gestational age-independent indices of fetal growth
Identification of the small for gestational age fetus with the use of gestational age-independent indices of fetal growth Michael Y. Divon, M.D., Paul F. Chamberlain, M.D., Louise Sipos, R.N., M.Ed., Frank A. Manning, M.D., and Lawrence D. Platt, M.D. Los Angeles, California, and Winnipeg, Manitoba, Canada This study reviews the roles of sonographic assessment of the rate of growth of the fetal abdominal Circumference, the f!;lmur length/abdominal circumference ratio, and qualitative determination of amniotic fluid volume as gestational age-independent indices for identification of the small for gestational age fetus. The sensitivity and specificity for single and combinations of test results were evaluated in 50 appropriate for gestational age and 40 small for gestational age fetuses. Positive and negative predictive values were derived for the general population. Our results indicate that either a rate of growth of the fetal abdominal circumference ,;;10 mm/14 days or a femur length/abdominal circumference ratio ;;.23.5 correctly identifies most small for gestational age fetuses. When the general population is screened, only 15% of small for gestational age fetuses will be missed by this combination of criteria. The presence of a pocket of amniotic fluid ,;;2.0 em is highly suggestive of small for gestational age fetus. However, the presence of a pocket of amniotic fluid >2.0 em does not guarantee an appropriate for gestational age fetus. We conclude that these gestational age-independent indices of fetal growth offer useful tools for differentiating between the small for gestational age and appropriate for gestational age fetus. (AM J 0BSTET GYNECOL 1986;155:1197-1201.)
a
Key words: Small for gestational age, fetal growth, amniotic fluid volume, femur length/ abdominal circumference ratio, ultrasound
The introduction of high-resolution real-time ultrasonograph into obstetric practice in recent years has provided an additional tool for antenatal detection of the small for gestational age fetus. With this diagnostic modality, it is now possible to assess a variety of fetal growth parameters [biparietal diameter, head circumference, abdominal: circumference, and femur length]. 1 Additionally, estimated fetal weight may be derived with the use of one or more of the above measurements.2 Growth curves throughout gestation for biparietal diameter, head circumference, abdominal circumference, and femur length are available as well as tables indicating estimated fetal weight at various gestational ages. 1 The use of these parameters for antenatal identification of the small for gestational age fetus with diagnostic real-time ultrasound has also been reported. The biparietal diameter, head circumference: abdominal circumference ratio, and estimated fetal weight may be used to this end if gestational age is
accurately known. 3-6 However, uncertainty about the gestational age occurs frequently and makes the differentiation between the appropriate for gestational age and the small for gestational age fetus difficult. 7 Gestational age-independent indices of fetal growth are therefore of value. In a recent report, Hadlock et a!. 8 reported that the femur length/abdominal circumference ratio is constant from 21 weeks' gestational age to term in the normally growing fetus. Similar results were reported by Jeanty et al. 9 These authors have suggested calculation of this ratio as a gestational age-independent index for the detection of asymmetric intrauterine growth retardation. Sonographic assessment of amniotic fluid volume (termed qualitative amniotic fluid volume determination) has also been suggested as a gestational age-independent method for the antepartum detection of intrauterine growth retardation. 10 This observation has recently been confirmed in a larger patient population. 11
From the Department of Obstetrics and Gynecology, University of Southern California School of Medicine and Women's Hospital, Los Angeles County/University of Southern California Medical Center, and the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology and Reproductive Sciences, University of Manitoba. Sponsored by the Society for Gynecologic Investigation. Reprint requests: Michael Y. Divon, M.D., Albert Einstein College of Medicine, Department of Obstetrics and Gynecology, Belfer Bldg., Room 501, 1300 Morris Park Ave., Bronx, NY 10461.
. Since growth of the fetal abdominal circumference appears to be linear from 15 weeks' gestational age onward, 12-14 determination of the rate of growth offers an additional, gestational age-independent index for identification of the small for gestational age fetus. The purpose of this study was to review the roles of sonographic assessment of the rate of growth of the fetal abdominal circumference, the femur length/ab-
1197
1198 Divon et al.
December 1986 Am J Obstet Gynecol
Table I. Indications for referral
APPROPRIATE FOR GESTATIONAL AGE
4
Group 1: Appropriate for gestational age
...z
U)
1
w
~ 6
a:
SMALL FOR GESTATIONAL AGE
~ 5 ::; :::>
z
4
-10
-5
10
15
20
25
30
35
RATE OF GROWTH OF FETAL ABDOMINAL CIRCUMFERENCE (mm/14 DAYS)
Fig. 1. Rate of growth of abdominal circumference in small
for gestational age and appropriate for gestational age fetuses. domina! circumference ratio, and the qualitative determination of amniotic fluid volume in identification of the small for gestational age fetus.
I
I
%
n
13
26
24
60
6 10 7 14 50
12 20 14 28 100
5 2
12.5 5
9
22.5 100
Indication
n
Suspected small for gestational age Hypertension Postterm Diabetes Other Total
;: 0
< 0.
Group 2: Small for gestational age
40
%
domina! circumference, femur length/abdominal circumference ratio, and qualitative determination of amniotic fluid volume for each group were determined. Results were compared with the use of Student's t test and considered significant if p values <0.05 were found. The sensitivity, specificity, and positive and negative predictive values (with the use of Bayes' theorem) for each test result alone and for combinations of test results were calculated. 16
Patients and methods
The study population consists of patients evaluated in the Fetal Assessment Units at the University of Manitoba between January 1982 and June 1983. Criteria for inclusion in the study were as follows: (1) two ultrasound examinations at least 14 days apart at which the abdominal circumference, femur length, and qualitative determination of amniotic fluid volume were recorded, (2) last sonographic examination within 5 days of delivery, (3) singleton pregnancy, and (4) intact membranes at the time of ultrasound examinations.' 5 Measurements were made by several experienced observers. The rate of growth of the fetal abdominal circumference was expressed in millimeters per 14 days and calculated as follows: rAC = (final AC - previous AC) X 14/(number of days between examinations) where r AC equals the rate of growth of the fetal abdominal circumference and previous AC equals abdominal circumference as determined at an examination at least 14 days before final abdominal circumference. With the use of a growth curve specific to this study population, the fetuses were divided into two groups depending on the percentile for gestational age and sex of their actual weight at birth [group 1 = appropriate for gestational age (;,.tenth to :s;ninetieth), group 2 = small for gestational age (
Results
Ninety patients seen during the study period fulfilled all the criteria and form the population for this report. These patients were referred for sonographic examination for a variety of reasons (Table I). Of these 90 patients, 50 (55.5%) were delivered of appropriate for gestational age fetuses (group 1) and 40 (44.5%) were delivered of small for gestational age fetuses (group 2). The mean actual birth weight was 3293 ± 393 gm ( ± SD) in group 1 and 2345 ± 468 gm in group 2. These results are significantly different (p < 0.01). The mean gestational age was 38.9 ± 1.7 weeks in group 1 and 38.5 ± 2.2 weeks in group 2. These results are not significantly different. The mean, standard deviation of the mean, and range of the rate of growth of the fetal abdominal circumference, femur length/abdominal circumference ratio, and qualitative determination of amniotic fluid volume in both the appropriate for gestational age and small for gestational age groups are shown in Table II. The mean rate of growth of the fetal abdominal circumference in the appropriate for gestational age and small for gestational age groups was 14.7 ± 7.1 and 6.0 ± 4.9 mm/14 days, respectively (p < 0.01). The distribution of rates of growth of abdominal circumference for each group is shown in Fig. 1. The mean femur length/abdominal circumference ratio in the appropriate for gestational age and small for gestational age groupswas21.85 ± 1.15and23.18 ± 1.7,respectively (p < 0.01). The mean qualitative determination of amniotic fluid volume in the appropriate for gestational
Gestational age-independent indices of fetal growth
Volume 155 Number 6
1199
Table II. Rate of growth of fetal abdominal circumference, femur length/abdominal circumference ratio, and amniotic fluid volume in appropriate for gestational age and small for gestational age fetuses Group 1: Appropriate for gestational age
Group 2: Small for gestational age
14.7 ± 7.1 2.5 to 35.5
6.0 ± 4.9 -4.8 to 18.0
p < 0.01
21.85 ± 1.15 18.8 to 24.2
23.18 ± 1.7 19.0 to 26.1
p < 0.01
Rate of growth of fetal abdominal circumference (mm/14 days) Mean± SD Range Femur length/abdominal circumference ratio Mean± SD Range Amniotic fluid volume (em) Mean± SD Range
5.44 ± 1.15 3.1 to 8.3
Significance
4.96 ± 1.48 1.2 to 7.8
NS
Table III. Sensitivity, specificity, and positive and negative predictive values for single and combinations of test results Test
Results
Sensitivity
Specificity
Positive predictive value* (%)
Negative predictive value* (%)
A B
1 2 3 1+ 2 1+ 3 2 + 3 1+ 2 + 3 Either 1 or 2 Either 1 or 2 or 3
0.85 0.55 0.10 0.40 0.10 0.10 0.10 0.90 0.90
0.74 0.90 1.0 0.74 1.0 1.0 1.0 0.74 0.74
26.6 38.0 100 14.5 100 100 100 28.0 28.0
2.2 5.5 9.0 8.5 9.0 9.0 9.0 1.5 1.5
c
D E F G H I
= Rate of growth of fetal abdominal circumference ~10 mm/14 days; 2 = femur length/abdominal circumference ratio ;.23.5; 3 = largest pocket of amniotic fluid (quantitative amniotic fluid volume determination) ~2.0 em (vertical depth). *With the use of Bayes' theorem, 16 positive predictive value = [(Se) (P)]/[(Se) (P) + (I - Sp) (I - P)], negative predictive value = 1 - [(Sp) (1 - p)]/[(sp) (I - p) + (1 - Se) (P)], where Se = sensitivity of test, Sp = specificity of test, and P = prevalence of disease in general population (10% for small for gestational age). age and small for gestational age groups was 5.44 ± 1.15 and 4.96 ± 1.48 em respectively (not significant). The sensitivity, specificity, and positive and negative predictive values for single and combinations of test results in identifying the small for gestational age fetus are shown in Table III.
Comment The results of this sonographic study indicate that there is a significant difference in the rate of growth of the fetal abdominal circumference between the small for gestational age and the appropriate for gestational age fetus. The mean rate of growth of the fetal abdominal circumference in the appropriate for gestational age fetus was 14.7 ± 7.1 mm/14 days ( ± SD) as compared with 6.0 ± 4.9 mm/14 days in the small for gestational age fetus (p < 0.01). The fact that measurements of abdominal circumference were made by several observers combined with the inherent inaccuracy in sonographic estimation of abdominal circumference may provide a partial explanation of the wide range and SD found in this study. It is of interest to note that a negative rate of growth of the fetal abdom-
ina! circumference was found in three of the 40 SGA fetuses (Fig. 1). This may represent either observer error in measurement of abdominal circumference or actual weight loss by the fetus in utero. Inspecting the results in Fig. 1, we suggest that a rate of growth of the fetal abdominal circumference ~ 10 mm/14 days is a reasonable compromise between sensitivity on the one hand and specificity on the other. Values below this rate would have identified 85% of the small for gestational age fetuses, while 74% of all appropriate for gestational age fetuses would have been identified by values above this rate. The positive predictive value for this cut-offlevel is 26.6%. This implies that in the general population patients with a rate of growth of the fetal abdominal circumference ~ 10 mml 14 days will have an increase in the probability of a small for gestational age fetus from 10% before testing ( 10% being the prevalence of small for gestational age fetuses in the general population) to 26.6% after testing. This relatively low positive predictive value could be improved by choosing a lower cutoff level (Fig. 1). However, when the test is used for screening, the focus should be on the predictive value of a negative test (that
1200 Divon et al.
is, chances of having the disease even though the test is negative). With a cutoff level ~10 mm/14 days, the predictive value of a negative test is 2.2%. This value implies that when the general population is screened for small for gestational age fetuses, patients with a rate of growth of the fetal abdominal circumference > 10 mml 14 days have a decrease in the probability of having a small for gestational age fetus from 10% before testing to a low 2.2% after testing. Normal values for the rate of growth of the fetal abdominal circumference were calculated from fetuses who were between the tenth and ninetieth percentiles of weight for gestation at delivery. These patients were referred for sonographic examination for a variety of reasons (Table 1). It must therefore be stressed that this sample is not truly representative of the general obstetric population. Nevertheless, comparison of the rates of growth we found to growth rates derived from previously published growth curves disclosed reasonable similarities. Hadlock et al. 14 have recently reported on the "fetal abdominal circumference as a predictor of menstrual age." Their data show that growth of the fetal abdominal circumference is linear from 15 weeks of gestation to term. Growth rates of 18 to 20 mm/14 days can be derived from these data. The rate of growth as described in this study (that is, rate of growth of the fetal abdominal circumference) does not require the knowledge of gestational age of the fetus beyond 15 weeks; therefore we suggest that it could be used as a gestational age-independent variable. This study has outlined the roles of several gestational age-independent indices of fetal growth (that is, rate of growth of the fetal abdominal circumference, femur length/abdominal circumference ratio, and qualitative determination of amniotic fluid volume) in identifying the small for gestational age fetus. The sensitivity, specificity, and positive and negative predictive values for single and combinations of test results have been presented. When single tests alone are considered, determination of the rate of growth of the abdominal circumference and assessment of the femur length/abdominal circumference ratio perform similarly in identification of the small for gestational age fetus. The lower negative predictive value for the rate of growth of abdominal circumference may favor it over the femur length/abdominal circumference ratio for screening purposes (Table III). The sensitivity and specificity of the femur length/abdominal circumference ratio in our study compare favorably with those reported by Hadlock et al. 8 These authors, using a femur length/ abdominal circumference ratio of >23.5 for identification of the fetus with intrauterine growth retardation, reported a sensitivity and specificity of 0.6 and 0.9 respectively. The corresponding figures from our study with the use of a femur length/abdominal circumfer-
December 1986 Am J Obstet Gynecol
ence ratio of ;;;.:23.5 for identification of the small for gestational age fetus were 0.55 and 0.9, respectively. The positive predictive power for an abnormal femur length/abdominal circumference ratio reported by these authors was 25%. 8 The positive predictive value for an abnormal femur length/abdominal circumference ratio of ;;;.:23.5 in our study was 38%. It is probable that the higher positive predictive value in our study is related to the prevalence of the condition being screened ( 10% for small for gestational age rather than 5% for intrauterine growth retardation, as in the study by Hadlock et al.). No significant difference in qualitative determination of amniotic fluid volume was noted between the appropriate for gestational age and small for gestational age groups. The specificity of 1.0 and negative predictive value of 9% found in this study compare favorably with similar values derived from the report of Chamberlain et al. 11 The corresponding figures in that study were 0.97 and 9%, respectively. In this study, the sensitivity of qualitative determination of amniotic fluid volume of ~2 em for identification of the small for gestational age fetus was low (0.1). This is in keeping with previous reports. Chamberlain et al." have reported a sensitivity of 0.13 for a qualitative determination of amniotic fluid volume of ~2 em in identification of the small for gestational age fetus. Philipson et al. 17 reported a sensitivity of 0.155 for oligohydramnios diagnosed sonographically in identification of the small for gestational age fetus. This low sensitivity is of importance as it indicates that the vast m~ority of small for gestational age fetuses have a "normal" volume of amniotic fluid. When combinations of test results are considered, the finding of either a rate of growth of the fetal abdominal circumference of ~ 10 mm/ 14 days or a femur length/ abdominal circumference ratio of ;;;.:23.5 probably represents the best balance between sensitivity, specificity, and positive and negative predictive values. A negative predictive value of 1.5% implies that patients who have either a rate of growth of the fetal abdominal circumference> 10 mm/14 days or a femur length/abdominal circumference ratio <23.5 have a decrease in the probability of having a small for gestational age fetus from 10% before testing to 1.5% after testing (Table III, H). This low negative predictive value makes it an attractive test combination for routine clinical purposes. The presence of a pocket of amniotic fluid ~2 em is highly suggestive of a small for gestational age fetus (100% positive predictive value in this study). However, the presence of a pocket of amniotic fluid >2 em does not guarantee a normal outcome (9% negative predictive value in this study). In conclusion, gestational age-independent indices of fetal growth offer an attractive tool for differentia-
Volume 155 Number 6
tion between the appropriate for gestational age and the small for gestational age fetus. Prospective studies to evaluate the role of these parameters in high-risk obstetric care are presently under way in our institutions. REFERENCES 1. Jeanty P, Romero R. Obstetrical ultrasound. New York: McGraw Hill, 1984. 2. Hadlock FP, Harrist RB, Carpenter RJ, Deter RL, Park SK. Sonographic estimation of fetal weight. Radiology 1984;150:535. 3. Deter RL, Harrist RB, Hadlock FP, Carpenter RJ. The use of ultrasound in the detection of intrauterine growth retardation: a review. JCU 1982; 10:9. 4. Ott WJ, Doyle S. Ultrasonic diagnosis of altered fetal growth by use of a normal ultrasonic fetal weight curve. Obstet Gynecol 1984;63:201. 5. Chervenak FA, Romero R, Berkowitz RL, Scott D, Tortora M, Hobbins JC. Use of sonographic estimated fetal weight in the prediction of intrauterine growth retardation. Am J Perinatal 1984; 1:298. 6. Sabbagaha RE, Hughey M, Depp R. The assignment of growth-adjusted sonographic age (GASA): a simplified method. Obstet Gynecol 1978;51 :383. 7. Campbell S. The assessment of fetal development by diagnostic ultrasound. Clin Perina to! 197 4; 1:507. 8. Hadlock FP, Deter RL, Harrist RB, Roecker E, Park SK. A date-independent predictor of intrauterine growth retardation: femur length: abdominal circumference ratio. Am] Roentgenol1983;141:979.
Gestational age-independent indices of fetal growth
1201
9. Jeanty P, Cousaert E, Cantraine F. Normal growth of the abdominal perimeter. Am] Perinatoll984;1:129. 10. Manning FA, Hill LM, Platt LD. Qualitative amniotic fluid determination by ultrasound: antepartum detection of intrauterine growth retardation. AM J OBSTET GYNECOL 1981;139:254. 11. Chamberlain PF, Manning FA, Morrison I, Harman CR, Lange IR. Ultrasound evaluation of amniotic fluid volume. I. The significance of marginal and decreased amniotic fluid volumes to perinatal outcome. AM J OBSTET GYNECOL 1984; 150:250. 12. Deter RL, Harrist RB, Hadlock FP, Carpenter RJ. Fetal head and abdominal circumferences. II. A critical re-evaluation of the relationship to menstrual age. JCU 1982;10:365. 13. Hobbins JC, Grannum PAT, Berkowitz RL, Silverman R, Mahoney MJ. Ultrasound in the diagnosis of congenital anomalies. AMj 0BSTET GYNECOL 1979; 134:331. 14. Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal abdominal circumference as a predictor of menstrual age. Am] Radiol1982;139:367. 15. Divon MY, Chamberlain PF, Sipos L, Platt LD. Underestimation of fetal weight in SPROM. J Ultrasound Med 1984;3:529. 16. Van De Merwe JP. Interpretation of a laboratory test using a programmable hand-held calculator. Calculation of posterior probability and relative risk on the basis of prior probability and sensitivity, specificity and results of the test. Comput Bioi Med 1984;14:107. 17. Philipson EH, Sokol RJ, Williams T. Oligohydramnios: clinical associations and predictive value for intrauterine growth retardation. AMJ 0BSTETGYNECOL 1983;146:271.
a
Key words: Small for gestational age, fetal growth, amniotic fluid volume, femur length/ abdominal circumference ratio, ultrasound
The introduction of high-resolution real-time ultrasonograph into obstetric practice in recent years has provided an additional tool for antenatal detection of the small for gestational age fetus. With this diagnostic modality, it is now possible to assess a variety of fetal growth parameters [biparietal diameter, head circumference, abdominal: circumference, and femur length]. 1 Additionally, estimated fetal weight may be derived with the use of one or more of the above measurements.2 Growth curves throughout gestation for biparietal diameter, head circumference, abdominal circumference, and femur length are available as well as tables indicating estimated fetal weight at various gestational ages. 1 The use of these parameters for antenatal identification of the small for gestational age fetus with diagnostic real-time ultrasound has also been reported. The biparietal diameter, head circumference: abdominal circumference ratio, and estimated fetal weight may be used to this end if gestational age is
accurately known. 3-6 However, uncertainty about the gestational age occurs frequently and makes the differentiation between the appropriate for gestational age and the small for gestational age fetus difficult. 7 Gestational age-independent indices of fetal growth are therefore of value. In a recent report, Hadlock et a!. 8 reported that the femur length/abdominal circumference ratio is constant from 21 weeks' gestational age to term in the normally growing fetus. Similar results were reported by Jeanty et al. 9 These authors have suggested calculation of this ratio as a gestational age-independent index for the detection of asymmetric intrauterine growth retardation. Sonographic assessment of amniotic fluid volume (termed qualitative amniotic fluid volume determination) has also been suggested as a gestational age-independent method for the antepartum detection of intrauterine growth retardation. 10 This observation has recently been confirmed in a larger patient population. 11
From the Department of Obstetrics and Gynecology, University of Southern California School of Medicine and Women's Hospital, Los Angeles County/University of Southern California Medical Center, and the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology and Reproductive Sciences, University of Manitoba. Sponsored by the Society for Gynecologic Investigation. Reprint requests: Michael Y. Divon, M.D., Albert Einstein College of Medicine, Department of Obstetrics and Gynecology, Belfer Bldg., Room 501, 1300 Morris Park Ave., Bronx, NY 10461.
. Since growth of the fetal abdominal circumference appears to be linear from 15 weeks' gestational age onward, 12-14 determination of the rate of growth offers an additional, gestational age-independent index for identification of the small for gestational age fetus. The purpose of this study was to review the roles of sonographic assessment of the rate of growth of the fetal abdominal circumference, the femur length/ab-
1197
1198 Divon et al.
December 1986 Am J Obstet Gynecol
Table I. Indications for referral
APPROPRIATE FOR GESTATIONAL AGE
4
Group 1: Appropriate for gestational age
...z
U)
1
w
~ 6
a:
SMALL FOR GESTATIONAL AGE
~ 5 ::; :::>
z
4
-10
-5
10
15
20
25
30
35
RATE OF GROWTH OF FETAL ABDOMINAL CIRCUMFERENCE (mm/14 DAYS)
Fig. 1. Rate of growth of abdominal circumference in small
for gestational age and appropriate for gestational age fetuses. domina! circumference ratio, and the qualitative determination of amniotic fluid volume in identification of the small for gestational age fetus.
I
I
%
n
13
26
24
60
6 10 7 14 50
12 20 14 28 100
5 2
12.5 5
9
22.5 100
Indication
n
Suspected small for gestational age Hypertension Postterm Diabetes Other Total
;: 0
< 0.
Group 2: Small for gestational age
40
%
domina! circumference, femur length/abdominal circumference ratio, and qualitative determination of amniotic fluid volume for each group were determined. Results were compared with the use of Student's t test and considered significant if p values <0.05 were found. The sensitivity, specificity, and positive and negative predictive values (with the use of Bayes' theorem) for each test result alone and for combinations of test results were calculated. 16
Patients and methods
The study population consists of patients evaluated in the Fetal Assessment Units at the University of Manitoba between January 1982 and June 1983. Criteria for inclusion in the study were as follows: (1) two ultrasound examinations at least 14 days apart at which the abdominal circumference, femur length, and qualitative determination of amniotic fluid volume were recorded, (2) last sonographic examination within 5 days of delivery, (3) singleton pregnancy, and (4) intact membranes at the time of ultrasound examinations.' 5 Measurements were made by several experienced observers. The rate of growth of the fetal abdominal circumference was expressed in millimeters per 14 days and calculated as follows: rAC = (final AC - previous AC) X 14/(number of days between examinations) where r AC equals the rate of growth of the fetal abdominal circumference and previous AC equals abdominal circumference as determined at an examination at least 14 days before final abdominal circumference. With the use of a growth curve specific to this study population, the fetuses were divided into two groups depending on the percentile for gestational age and sex of their actual weight at birth [group 1 = appropriate for gestational age (;,.tenth to :s;ninetieth), group 2 = small for gestational age (
Results
Ninety patients seen during the study period fulfilled all the criteria and form the population for this report. These patients were referred for sonographic examination for a variety of reasons (Table I). Of these 90 patients, 50 (55.5%) were delivered of appropriate for gestational age fetuses (group 1) and 40 (44.5%) were delivered of small for gestational age fetuses (group 2). The mean actual birth weight was 3293 ± 393 gm ( ± SD) in group 1 and 2345 ± 468 gm in group 2. These results are significantly different (p < 0.01). The mean gestational age was 38.9 ± 1.7 weeks in group 1 and 38.5 ± 2.2 weeks in group 2. These results are not significantly different. The mean, standard deviation of the mean, and range of the rate of growth of the fetal abdominal circumference, femur length/abdominal circumference ratio, and qualitative determination of amniotic fluid volume in both the appropriate for gestational age and small for gestational age groups are shown in Table II. The mean rate of growth of the fetal abdominal circumference in the appropriate for gestational age and small for gestational age groups was 14.7 ± 7.1 and 6.0 ± 4.9 mm/14 days, respectively (p < 0.01). The distribution of rates of growth of abdominal circumference for each group is shown in Fig. 1. The mean femur length/abdominal circumference ratio in the appropriate for gestational age and small for gestational age groupswas21.85 ± 1.15and23.18 ± 1.7,respectively (p < 0.01). The mean qualitative determination of amniotic fluid volume in the appropriate for gestational
Gestational age-independent indices of fetal growth
Volume 155 Number 6
1199
Table II. Rate of growth of fetal abdominal circumference, femur length/abdominal circumference ratio, and amniotic fluid volume in appropriate for gestational age and small for gestational age fetuses Group 1: Appropriate for gestational age
Group 2: Small for gestational age
14.7 ± 7.1 2.5 to 35.5
6.0 ± 4.9 -4.8 to 18.0
p < 0.01
21.85 ± 1.15 18.8 to 24.2
23.18 ± 1.7 19.0 to 26.1
p < 0.01
Rate of growth of fetal abdominal circumference (mm/14 days) Mean± SD Range Femur length/abdominal circumference ratio Mean± SD Range Amniotic fluid volume (em) Mean± SD Range
5.44 ± 1.15 3.1 to 8.3
Significance
4.96 ± 1.48 1.2 to 7.8
NS
Table III. Sensitivity, specificity, and positive and negative predictive values for single and combinations of test results Test
Results
Sensitivity
Specificity
Positive predictive value* (%)
Negative predictive value* (%)
A B
1 2 3 1+ 2 1+ 3 2 + 3 1+ 2 + 3 Either 1 or 2 Either 1 or 2 or 3
0.85 0.55 0.10 0.40 0.10 0.10 0.10 0.90 0.90
0.74 0.90 1.0 0.74 1.0 1.0 1.0 0.74 0.74
26.6 38.0 100 14.5 100 100 100 28.0 28.0
2.2 5.5 9.0 8.5 9.0 9.0 9.0 1.5 1.5
c
D E F G H I
= Rate of growth of fetal abdominal circumference ~10 mm/14 days; 2 = femur length/abdominal circumference ratio ;.23.5; 3 = largest pocket of amniotic fluid (quantitative amniotic fluid volume determination) ~2.0 em (vertical depth). *With the use of Bayes' theorem, 16 positive predictive value = [(Se) (P)]/[(Se) (P) + (I - Sp) (I - P)], negative predictive value = 1 - [(Sp) (1 - p)]/[(sp) (I - p) + (1 - Se) (P)], where Se = sensitivity of test, Sp = specificity of test, and P = prevalence of disease in general population (10% for small for gestational age). age and small for gestational age groups was 5.44 ± 1.15 and 4.96 ± 1.48 em respectively (not significant). The sensitivity, specificity, and positive and negative predictive values for single and combinations of test results in identifying the small for gestational age fetus are shown in Table III.
Comment The results of this sonographic study indicate that there is a significant difference in the rate of growth of the fetal abdominal circumference between the small for gestational age and the appropriate for gestational age fetus. The mean rate of growth of the fetal abdominal circumference in the appropriate for gestational age fetus was 14.7 ± 7.1 mm/14 days ( ± SD) as compared with 6.0 ± 4.9 mm/14 days in the small for gestational age fetus (p < 0.01). The fact that measurements of abdominal circumference were made by several observers combined with the inherent inaccuracy in sonographic estimation of abdominal circumference may provide a partial explanation of the wide range and SD found in this study. It is of interest to note that a negative rate of growth of the fetal abdom-
ina! circumference was found in three of the 40 SGA fetuses (Fig. 1). This may represent either observer error in measurement of abdominal circumference or actual weight loss by the fetus in utero. Inspecting the results in Fig. 1, we suggest that a rate of growth of the fetal abdominal circumference ~ 10 mm/14 days is a reasonable compromise between sensitivity on the one hand and specificity on the other. Values below this rate would have identified 85% of the small for gestational age fetuses, while 74% of all appropriate for gestational age fetuses would have been identified by values above this rate. The positive predictive value for this cut-offlevel is 26.6%. This implies that in the general population patients with a rate of growth of the fetal abdominal circumference ~ 10 mml 14 days will have an increase in the probability of a small for gestational age fetus from 10% before testing ( 10% being the prevalence of small for gestational age fetuses in the general population) to 26.6% after testing. This relatively low positive predictive value could be improved by choosing a lower cutoff level (Fig. 1). However, when the test is used for screening, the focus should be on the predictive value of a negative test (that
1200 Divon et al.
is, chances of having the disease even though the test is negative). With a cutoff level ~10 mm/14 days, the predictive value of a negative test is 2.2%. This value implies that when the general population is screened for small for gestational age fetuses, patients with a rate of growth of the fetal abdominal circumference > 10 mml 14 days have a decrease in the probability of having a small for gestational age fetus from 10% before testing to a low 2.2% after testing. Normal values for the rate of growth of the fetal abdominal circumference were calculated from fetuses who were between the tenth and ninetieth percentiles of weight for gestation at delivery. These patients were referred for sonographic examination for a variety of reasons (Table 1). It must therefore be stressed that this sample is not truly representative of the general obstetric population. Nevertheless, comparison of the rates of growth we found to growth rates derived from previously published growth curves disclosed reasonable similarities. Hadlock et al. 14 have recently reported on the "fetal abdominal circumference as a predictor of menstrual age." Their data show that growth of the fetal abdominal circumference is linear from 15 weeks of gestation to term. Growth rates of 18 to 20 mm/14 days can be derived from these data. The rate of growth as described in this study (that is, rate of growth of the fetal abdominal circumference) does not require the knowledge of gestational age of the fetus beyond 15 weeks; therefore we suggest that it could be used as a gestational age-independent variable. This study has outlined the roles of several gestational age-independent indices of fetal growth (that is, rate of growth of the fetal abdominal circumference, femur length/abdominal circumference ratio, and qualitative determination of amniotic fluid volume) in identifying the small for gestational age fetus. The sensitivity, specificity, and positive and negative predictive values for single and combinations of test results have been presented. When single tests alone are considered, determination of the rate of growth of the abdominal circumference and assessment of the femur length/abdominal circumference ratio perform similarly in identification of the small for gestational age fetus. The lower negative predictive value for the rate of growth of abdominal circumference may favor it over the femur length/abdominal circumference ratio for screening purposes (Table III). The sensitivity and specificity of the femur length/abdominal circumference ratio in our study compare favorably with those reported by Hadlock et al. 8 These authors, using a femur length/ abdominal circumference ratio of >23.5 for identification of the fetus with intrauterine growth retardation, reported a sensitivity and specificity of 0.6 and 0.9 respectively. The corresponding figures from our study with the use of a femur length/abdominal circumfer-
December 1986 Am J Obstet Gynecol
ence ratio of ;;;.:23.5 for identification of the small for gestational age fetus were 0.55 and 0.9, respectively. The positive predictive power for an abnormal femur length/abdominal circumference ratio reported by these authors was 25%. 8 The positive predictive value for an abnormal femur length/abdominal circumference ratio of ;;;.:23.5 in our study was 38%. It is probable that the higher positive predictive value in our study is related to the prevalence of the condition being screened ( 10% for small for gestational age rather than 5% for intrauterine growth retardation, as in the study by Hadlock et al.). No significant difference in qualitative determination of amniotic fluid volume was noted between the appropriate for gestational age and small for gestational age groups. The specificity of 1.0 and negative predictive value of 9% found in this study compare favorably with similar values derived from the report of Chamberlain et al. 11 The corresponding figures in that study were 0.97 and 9%, respectively. In this study, the sensitivity of qualitative determination of amniotic fluid volume of ~2 em for identification of the small for gestational age fetus was low (0.1). This is in keeping with previous reports. Chamberlain et al." have reported a sensitivity of 0.13 for a qualitative determination of amniotic fluid volume of ~2 em in identification of the small for gestational age fetus. Philipson et al. 17 reported a sensitivity of 0.155 for oligohydramnios diagnosed sonographically in identification of the small for gestational age fetus. This low sensitivity is of importance as it indicates that the vast m~ority of small for gestational age fetuses have a "normal" volume of amniotic fluid. When combinations of test results are considered, the finding of either a rate of growth of the fetal abdominal circumference of ~ 10 mm/ 14 days or a femur length/ abdominal circumference ratio of ;;;.:23.5 probably represents the best balance between sensitivity, specificity, and positive and negative predictive values. A negative predictive value of 1.5% implies that patients who have either a rate of growth of the fetal abdominal circumference> 10 mm/14 days or a femur length/abdominal circumference ratio <23.5 have a decrease in the probability of having a small for gestational age fetus from 10% before testing to 1.5% after testing (Table III, H). This low negative predictive value makes it an attractive test combination for routine clinical purposes. The presence of a pocket of amniotic fluid ~2 em is highly suggestive of a small for gestational age fetus (100% positive predictive value in this study). However, the presence of a pocket of amniotic fluid >2 em does not guarantee a normal outcome (9% negative predictive value in this study). In conclusion, gestational age-independent indices of fetal growth offer an attractive tool for differentia-
Volume 155 Number 6
tion between the appropriate for gestational age and the small for gestational age fetus. Prospective studies to evaluate the role of these parameters in high-risk obstetric care are presently under way in our institutions. REFERENCES 1. Jeanty P, Romero R. Obstetrical ultrasound. New York: McGraw Hill, 1984. 2. Hadlock FP, Harrist RB, Carpenter RJ, Deter RL, Park SK. Sonographic estimation of fetal weight. Radiology 1984;150:535. 3. Deter RL, Harrist RB, Hadlock FP, Carpenter RJ. The use of ultrasound in the detection of intrauterine growth retardation: a review. JCU 1982; 10:9. 4. Ott WJ, Doyle S. Ultrasonic diagnosis of altered fetal growth by use of a normal ultrasonic fetal weight curve. Obstet Gynecol 1984;63:201. 5. Chervenak FA, Romero R, Berkowitz RL, Scott D, Tortora M, Hobbins JC. Use of sonographic estimated fetal weight in the prediction of intrauterine growth retardation. Am J Perinatal 1984; 1:298. 6. Sabbagaha RE, Hughey M, Depp R. The assignment of growth-adjusted sonographic age (GASA): a simplified method. Obstet Gynecol 1978;51 :383. 7. Campbell S. The assessment of fetal development by diagnostic ultrasound. Clin Perina to! 197 4; 1:507. 8. Hadlock FP, Deter RL, Harrist RB, Roecker E, Park SK. A date-independent predictor of intrauterine growth retardation: femur length: abdominal circumference ratio. Am] Roentgenol1983;141:979.
Gestational age-independent indices of fetal growth
1201
9. Jeanty P, Cousaert E, Cantraine F. Normal growth of the abdominal perimeter. Am] Perinatoll984;1:129. 10. Manning FA, Hill LM, Platt LD. Qualitative amniotic fluid determination by ultrasound: antepartum detection of intrauterine growth retardation. AM J OBSTET GYNECOL 1981;139:254. 11. Chamberlain PF, Manning FA, Morrison I, Harman CR, Lange IR. Ultrasound evaluation of amniotic fluid volume. I. The significance of marginal and decreased amniotic fluid volumes to perinatal outcome. AM J OBSTET GYNECOL 1984; 150:250. 12. Deter RL, Harrist RB, Hadlock FP, Carpenter RJ. Fetal head and abdominal circumferences. II. A critical re-evaluation of the relationship to menstrual age. JCU 1982;10:365. 13. Hobbins JC, Grannum PAT, Berkowitz RL, Silverman R, Mahoney MJ. Ultrasound in the diagnosis of congenital anomalies. AMj 0BSTET GYNECOL 1979; 134:331. 14. Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal abdominal circumference as a predictor of menstrual age. Am] Radiol1982;139:367. 15. Divon MY, Chamberlain PF, Sipos L, Platt LD. Underestimation of fetal weight in SPROM. J Ultrasound Med 1984;3:529. 16. Van De Merwe JP. Interpretation of a laboratory test using a programmable hand-held calculator. Calculation of posterior probability and relative risk on the basis of prior probability and sensitivity, specificity and results of the test. Comput Bioi Med 1984;14:107. 17. Philipson EH, Sokol RJ, Williams T. Oligohydramnios: clinical associations and predictive value for intrauterine growth retardation. AMJ 0BSTETGYNECOL 1983;146:271.