British Journal of Obstetrics and Gynaecology March 1977. Vol 84. pp 1615-174
ULTRASOU ND MEASUREMENT OF THE FETAL HEAD TO ABDOMEN CIRCUMFERENCE RATIO IN THE ASSESSMENT OF GROWTH RETARDATION BY
S . CAMPBELL, Professor * AND
ALISON THOMS, Research Assistant * Institute of Obstetrics and Gynaecology, Queen Charlotte’s Hospital, London, W6
Summary Fetal head and abdomen circumference measurements have been made using pulsed ultrasound techniques. The mean head to abdomen (H/A) circumference ratio with 95 per cent confidence limits was determined in 568 normal pregnancies from 17 to 41 weeks menstrual age. The mean ratio was 1-18 at 17 weeks, but decreased slow1,y until 29 weeks when the ratio was 1 .I1 ; thereafter there was a sharp fall in the mean ratio to I .01 at 36 weeks and 0.96 at 40 weeks. The H/A circumference ratio was also determined in 3 1 small-for-dates fetuses within one week of delivery. In all cases the fetal weight as predicted from the fetal abdomen circumference measurement was below the 5th centile weight for gestation. The ratio was above the 95th centile limit in 22 (71 per cent) of these fetuses and evidence is presented to suggest that the H/A circumference ratio can be used to distinguish between symmetrical and asymmetrical growth retardation.
to rationalize some of the more recent therapeutic regimens designed to improve fetal growth (Bonnar et al, 1975; Varma and Curzen, 1973; Brettes et al, 1976) and would permit a more realistic appraisal of the short-term and longterm risks of the small-for-dates fetus, thus helping to decide the optimal time for delivery. From serial ultrasonic cephalometry studies, Campbell (1974~) described the low profile and the late flattening growth retardation patterns of the fetal biparietal diameter which would appear to conform to the symmetrical and asymmetrical growth retardation categories. It is likely, however, that more information would be gained about the type of growth retardation if the relationship between fetal head and body measurements could be studied. Asymmetrical growth retardation is manifested by a high fetal brain to liver ratio (Gruenwald, 1966) and while these organs cannot be directly measured, ultrasound circumference measure-
IT is now recognized that small-for-dates babies are not a homogenous population and that at least two morphological groups can be distinguished: one in whiich there is symmetrical reduction in the size of all organs (symmetrical growth retardation) and the second in which the baby has a long wasted body and a relatively large brain which has been preferentially protected from the full effects of the growth retarding stimulus (asymmetrical growth retardation). The aet iological mechanisms, perinatal risks and long term prognosis appear to differ between the two groups (Campbell, 1974a) and it would tie of great value if a reliable antenatal means of distinguishing between these types of growth retardation could be found. For example, it would help
* Present address : Department of Obstetrics and Gynaecology. King’s College Hospital, Denmark Hill, London SE5 165
166
CAMPBELL AND THOMS
ment of the fetal skull and the fetal abdomen at the level of the liver can be performed and would appear to be a logical alternative. N o information on the fetal head to abdomen (H/A) circumference ratio is at present available and the following study was designed to establish whether the H/A ratio was of clinical value in distinguishing between symmetrical and asymmetrical growth retardation.
METHODS Circumference measurements of the fetal head and abdomen were made from B-scan echograms obtained with the Diasonograph 4102 (Nuclear Enterprises Ltd, Edinburgh) using a frequency of 2 - 5 mHz and a velocity setting of 1540 metres per second. Measurements of the biparietal diameter (BPD) were also made by the combined A and B-scan method of Campbell (1968) using a sound velocity of 1600 metres per second (Willocks et al, 1964). Head circumference The method of head circumference measurement has not previously been described. The aim of the technique was to obtain a horizontal section of the fetal head which included both
the biparietal diameter (coronal plane) and the occipito-frontal diameter (saggital plane). The first part of the procedure involved making longitudinal scans of the fetal body to determine on which side the fetal spine was positioned; this manoeuvre was of importance later when determining the degree of flexion of the fetal head on the spine. The second part of the procedure was identical to that described for measurement of the fetal BPD (Campbell, 1968) in which longitudinal scans were made to determine the angle of inclination of the fetal head to the vertical axis. Subsequently, transverse scans were made with the transducer tilted to this angle so that a horizontal section of the fetal head was obtained; this was recognized by the appearance of the midline echo and the widest fetal head diameter at right angles to the midline echo was the biparietal diameter. Due to flexion of the fetal head, however, this transverse scan usually displayed the suboccipitobregmatic diameter in the sagittal plane. The final part of the procedure therefore involved rotating the scanning gantry (the direction of rotation being determined by the position of the fetal spine) until the head appeared as an ovoid and echoes from the third ventricle were
FIG.1 Echogram showing a horizontal section of a fetal head suitable for circumference measurement. The third ventricle is clearly shown in the midline one third of the distance from the synciput.
ULTRASONIC ASSESSMENT OF FETAL GROWTH RETARDATION
167
detected in the midline, one-third of the distance from the synciput (Fig. 1). This ensured that the occipito-frontal heald circumference section had been achieved.
known pregnancy complications and all babies, when delivered at term, were considered by paediatric assessment to be of appropriate neurological development.
Abdomen circumference The method of abdomen circumference measurement has been described previously (Campbell and Wilkin, 1975). The section was a transverse one through the liver at the level of the umbilical vein.
Fetal growth retardation There were 30 patients who were delivered of a total of 31 small-for-dates babies which were studied antenatally with H/A circumference ratio measurements. A small-for-dates baby was defined as being on or below the 5th centile limit according to the tables of Thompson et a1 (1968), corrections being made for maternal parity and the sex of the child. In these cases H/A circumference ratio studies were made for one or more of the following reasons: (i) a small-for-dates uterus as assessed by palpation, (ii) the occurrence of a complication which might prejudice placental function, and (iii) a fall below the 5th centile in the growth rate of the biparietal diameter measured serially (Campbell and Newman, 1971). In all cases, the fetal weight as predicted from the fetal abdomen circumference measurement (Campbell and Wilkin, 1975) was considered to be on or below the 5th centile weight for gestation (Thompson et al, 1968).
Measurement When the required sections of the head and abdomen were obtained, circumference measurements were made from a Polaroid photograph by means of a map measurer. As the HewlettPackard 141 B-oscilloscope incorporated in the Diasonograph could not contain life sized (5/5) images in the third1 trimester, 3/5 and 4/5 size images were measured, appropriate corrections being made. The usual reduction in image size of 0.93 to 1, which is a feature of this oscilloscope, was compensated for by altering the focal distance of the camera.
PATIE~NTS The study was arranged in two parts. First, 568 measurements were obtained from 523 patients in normal pregnancy to establish the mean, 5th and 95th per cent confidence limits for the fetal H/A circumj'erence ratio in relation to menstrual age. Second, measurement of this ratio was then made in pregnancies in which fetal growth retardation was suspected. Normal pregnancy Our data consisted of 568 measurements obtained from 523 patients with normal singleton pregnancies. The pregnancy was considered to be normal when the patient was sure of the date of her last menstrual period and had regular 28-day cycles. In most cases, the fetal age had also been determined by ultrasonic measurement of the embryonic crown-rump length up to 12 weeks amenorrhoea (Robinson, 1973) or fetal biparietal diameter measurement between 12 and 20 weeks (Campbell, 19743) and these assessments were in agreement with the menstrual age as calculated from the last menstrual period. In addition, there were no
,
,
n = 568
30.8
z0 . 7I6
20
24
28
32
36
40
MENSTRUAL AGE (weeks)
FIG.2 Graphic representation of data in Table I showing mean H/A circumference ratios with 5th and 95th centile confidence limits from 17 to 42 weeks menstrual age.
168
CAMPBELL AND THOMS
RESULTS Normal data The mean, 5th and 95th centile limits for the H/A circumference ratio were calculated (Table I) and plotted (Fig. 2) in 2-weekly groups to smooth out fluctuations due to small numbers. The mean H/A circumference ratio was 1 .18 at 17 weeks and decreased slowly until 29 weeks when the mean ratio was 1 .11; thereafter there was a sharp fall in the mean
ratio to 1.01 at 36 weeks and 0.96 at 40 weeks. The range between the confidence limits remained fairly constant at 0.18. Fetal growth retardation All the 31 small-for-dates fetuses had H/A circumference ratio measurements made within seven days of delivery. In each case the ratio was above the mean of the normal range (Fig. 3 and Table 11).
TABLE I Mean fetal HIA circumference ratios with 5th and 95th centile limits related to menstrual age from 13 to 42 weeks; values have been combined in 2-weekly groupings to smooth out fluctuations due to small numbers (568 individual measurements)
___
_ _ ~ ~
~~
~
~
_
H/A circumference ratio Menstrual age (weeks)
13-14 15-16 17-18 19-20 21-22 23-24 25-26 27-28 29-30 31-32 33-34 35-36 37-38 39-40 4142
Number Of measurements
__
18 39 77 54 41 22 18 36 23 31 42 49 67 41 4
5th centile
Mean
95th centile
1.14 1.05 1.07 1.09 1.06 1.05 1.04 1.05
1.23 1.22 1.18 1.18 1.15 1.13 1.13 1.13
0-99 0.96 0.96 0.93 0.92 0.87 0.93
1.07 1.04 1.02 0.98 0.97 0.96
1.31 1.39 1.29 1.26 1.25 1.21 1.22 1.22 1.21 1.17 1.11 1.11 1.05 1.06 1 .oo
1
.lo
TABLE I1 Summary of clinical details of the 31 growth retarded fetuses related to whether the HIA circumference ratio was normal or abnormal
H/A circumference ratio above 95th centile (22infants, 21 pregnancies) No. Severe Hypertension Forceps delivery Emergency Caesarean section Elective Caesarean section Intrapartum fetal distress Apgar score of 5 or less Congenital fetal abnormality Perinatal mortality
5 5 4 4 9
lot
23 23 18 18 53*1 48
0
-
2
9
* Percentage of patients who were allowed to labour.
t Stillbirths excluded.
Per cent
H/A circumference ratio above mean but within normal range (9 infants, 9 pregnancies) No.
Per cent
1
11
0 1
1 1 3t 2 1
11 11 14*t 31 22
11
_
ULTRASONIC ASSESSMENT OF FETAL GROWTH RETARDATION
\
169
Most of the 22 babies were described as wasted on paediatric examination. Six babies had thicknesses of eight skin-folds measured according to the method of Whitelaw (1976); the sum of the skin-fold thicknesses was below
Y
L,
0.7
Ljlllllllllllll
lb
20
24
28
32
36
Patient C H
MENSTRUAL ACE ( w e e k s ) FIG.3 H/A circumference ratios of 31 small-for-dates fetuses related to the normal range and classified according to the type of BPD growth curve. All measurements taken = late retardation of within 7 days of delivery. BPD growth; o = low growth profile for serial BPD measurenients; 0- == low B:PD growth profile with terminal arrest of growth; X = normal BPD increase.
Ratios above the 95th centile In this group of 21 patients (22 fetuses) there was one intrauterine death and one neonatal death. Both babies were examined postmortem and death was thought to be due to growth retardation and asphyxia. Five patients had severe hypertension (a blood pressure of 160/100 mm IHg or more) and the clinical diagnosis in all five was pre-eclampsia. Four patients required emergency Caesarean section (see below), a lfurther five required delivery by forceps and one patient had an assisted breech delivery. Four patients had an elective Caesarean section to protect the fetus from the stress of labour. Seven patients (including the one with the twin pregnancy) had normal vaginal deliveries. Of the 17 patients allowed to labour, 9 developed fetal distress as defined by late deceleration patterns on a cardiotocograph or a fetal scalp blood pH of less than 7.25. Ten of the 21 liveborn babies in this group had an Apgar score of 5 or less at one minute.
- Twins
&rth weiqhtr 1.92 L I 73 kq a t 37 weeks Apqarr 8 in I ~ 7 i In
40
I6
20
24
28
32
36
40
MENSTRUAL AGE ( w e e k s )
201-1 MENSTRUAL AGE IN W t E K S
FIG.4 Elevated fetal H/A circumference ratios in a twin pregnancy. Both BPD growth curves were of the ‘late flattening’ type.
170
CAMPBELL AND THOMS
the lower limit of normal (Whitelaw, unpublished data) in all cases. Seventeen fetuses showed normal growth of the BPD until the end of the second trimester or later and then developed a sharp slowing of the growth rate, thus conforming to the late flattening growth pattern (see Fig. 4). One of them showed some catch-up growth of the biparietal diameter after hospital admission (Fig. 5). Two of the 22 fetuses had a normal growth pattern on serial BPD measurement and the results obtained in one of these patients is shown in Figure 6. One patient showed a low BPD growth profile pattern followed by catch-up growth. Three patients had a low BPD growth profile initially and then developed late cessation of growth; two of these mothers smoked 15 cigarettes per day throughout pregnancy (Fig. 7).
Ratios between the mean and 95th centile In the nine patients with H/A circumference ratios within this range, there was one emergency Caesarean section but no forceps deliveries for fetal distress in labour. There was one elective Caesarean section because of concern for the condition of the fetus. Of the 8 patients allowed to labour, one developed fetal distress and three of the nine infants born had an Apgar score of 5 or less at one minute. Two babies had congenital defects: one had cystic fibrosis and the other had congenital rubella; these babies were not described as wasted at paediatric examination. Two fetuses had skin-fold measurements performed (Whitelaw, 1976) with values about the mean for the gestational age. Six of the nine fetuses had a persistent low BPD growth rate from early in the second trimester conforming to the low growth profile (Fig. 8). One of these six fetuses died in utero at 29 weeks gestation. The cause of death was a concealed antepartum haemorrhage. The pathologist described the placenta as abnormally small and noted that growth retardation was present. Three of the nine fetuses showed late growth retardation; one was of a very minor degree while the mother of another developed sudden late onset
pre-eclampsia and was delivered by Caesarean section.
Patient B K Forceps del. for fetal distress Birth weiqht 2 . 2 kq a t 37 weeks
0 7
I6
20
24 28 32 36 MENSTRUAL ACE (weeks)
40
I lor
-111111111111 Ib
22 26 30 34 MENSTRUAL AGE IN WEEKS
18
38
FIG.5 Elevated fetal H/A ratios in a fetus whose BPD growth curve showed late retardation followed by ‘catch-up’ growth.
171
ULTRASONIC ASSESSMENT OF FETAL GROWTH RETARDATION
Patient S P
Patient C B
Birth weiqht 1.17 kq at
Birth weiqht I 76 kq
36 weeks
Apqar
0 7
-5
in 1
0.7
20 24 28 32 36 MENSTRUAL AGE (weeks)
I6
40
I6
20 24 28 32 36 MENSTRUAL AGE (weeks)
40
i l L1-UIh
18
22
26
30
34
38
MENSTRUAL. AGE IN WEEKS
FIG.6 Elevated fetal H/A circumference ratios associated with normal growth of the BPD.
15
18 22 2b 30 34 M?NSTRUAL ACE IN W F F K S
38
FIG.7 Elevated fetal H/A circumference ratios in a patient who was a heavy smoker. The BPD growth curve showed early retardation with terminal cessation.
172
CAMPBELL AND THOMS
Patient A D Normal vaqinal delivery B i r t h weiqht 2.62 kq at 4 0 weeks Apqar 9 in I
0.7 I
I6
I
20
I
I
I
I
I
I
1
1
24 28 32 36 MENSTRUAL AGE (weeks)
I
I
I
40
201-1 MENSTRUAL AGE IN WEEKS
FIG.8 Normal fetal H/A circumference ratio associated with symmetrical growth retardation. The BPD growth curve showed the ‘low profile’ pattern.
DISCUSSION While accepting that measurement of the BPD by the combined A and B-scan technique
is the most accurate and most reproducible fetal measurement which can be obtained antenatally (Campbell, 1973), we decided t o use the fetal head circumference measurement when determining the ratio between head and body. This was principally because the BPD is only a single dimension of the head and is frequently not truly representative of the total fetal head and (by inference) brain size. This is especially true in dolichocephaly when the narrow BPD gives a false impression of subnormal head size. We believe the head circumference measurement to be more representative of brain size and this is supported by our finding that normal weight dolichocephalic fetuses have head circumference measurements within the normal range. Furthermore, postnatal studies show a good correlation between head circumference measurements and brain development (Bray et al, 1969; Dobbing, 1970). Only Levi and Erbsman (1975) have previously discussed antenatal ultrasonic measurement of the head circumference but from their description of the technique and the illustrative echogram it is doubtful whether the occipitofrontal diameter was always obtained in the sagittal plane. For reproducible and accurate measurement of the head circumference, we believe that it is essential to make allowance for flexion of the fetal head and the position of the third ventricle in the midline makes a very efficient indicator as to whether or not the optimal section has been obtained. We have found the mean standard deviation of three independent head circumference measurements in the third trimester to be 1 .X mm (Campbell, 1976) which represents a 1 .O per cent error; therefore, although reproducibility of head circumference measurements is not as precise as that of biparietal cephalometry, we believe that the advantages of using the circumference measurement outweigh this small loss of precision. The advantages of measuring the fetal abdomen rather than the fetal thorax have been discussed by Campbell and Wilkin (1975) and Campbell (1976). In particular, we believe that by using the umbilical vein as a reference point the reproducibility of abdomen circumference measurements is improved ; the mean standard deviation of three
ULTRASONIC ASSESSMENT OF FETAL GROWTH RETARDATION
independent fetal abdomen circumference measurements in our laboratory is 2.95 mm (Campbell, 1976) which represents an acceptable 2 . 0 per cent error. In addition, sections at the level of the umbilical vein and fetal liver would seem to be particularly appropriate in studies on the small-for-dates fetus, for the fetal liver is the most affected organ in asymmetrical growth retardation. Our normal data graph shows a changing H/A circumference raticr from 17 to 41 weeks menstrual age. At 17 weeks the mean head circumference measurement was I 8 per cent larger than that of the abdomen, but between 17 and 30 weeks there was only a gradual fall in the ratio as bolh head and abdomen measurements were increasing rapidly. Between 30 and 36 weeks there was a sharp reduction in the mean ratio due to slowing of head growth which was not demonstrated by the abdomen circumference measurement so that during the 36th week the mean head and abdomen circumference measurements were equal. Although there was some deceleration of abdomen circumference growth from 36 weeks to term, slowing down of head circumference growth was even more pronounced so that in late pregnancy the normally nourished fetus usually had a head to abdomen circumference ratio of less than 1. It is interesting to note that although the range for both the head and abdomen circumference measurements widened towards term (Campbell and Thorns, unpublished data) the confidence limits for the head to abdomen circumference ratio remained constant. By dividing the 31 small-for-dates fetuses in our study into two groups, dependent on whether the H/A circumference ratio was above or below the 95 per cent confidence limit, we were able to detect clinical differences between the groups. The wasted baby is typically caused by utero-placental vascular insufficiency and the clinical features of the 22 infants (21 pregnancies) which showed elevated ratios suggested this mechanism; there was a high incidence of severe pre-eclampsia, fetal distress in labour, operative intervention and low Apgar scores (Table 11) while all six babies appropriately investigated had significantly reduced skin-fold thicknesses. The aetiological factors associated
173
with symmetrical growth retardation are more complex but it is known that chronic maternal undernutrition (Rosso and Winick, 1974) heavy maternal smoking, fetal infection and genetic or chromosome abnormalities (Campbell, 19740) can all produce the stunted baby; if none of these factors can be implicated then an immunological mechanism may be at work (Billington, 1975). Whatever the mechanism, utero-placental vascular insufficiency is less commonly associated with this type of growth retardation and the nine small-for-dates fetuses in our study who had normal H/A circumference ratios had a lower incidence of maternal hypertension, fetal distress and operative intervention (Table 11). Furthermore, two babies in this group who had skin-fold thickness studies made showed no evidence of wasting. The two infants with congenital defects had ratios below the 95th centile limit and the possibility of the stunted abnormal baby should always be considered when a small fetus with a normal H/A circumference ratio is diagnosed antenatally. The good correlation between the two H/A circumference ratio groups and the type of cephalometry growth retardation pattern may be of importance in the long-term assessment of the small-for-dates infant. Fancourt et al ( 1 976), in follow-up studies of small-for-dates babies who were measured antenatally by serial cephalometry and who were examined at a mean age of 4 years, have shown that the fetus with the longest period of growth restriction (the low growth profile group) retained the somatic deficit in height, weight and (to a lesser extent) head circumference. Furthermore, those infants, who as fetuses showed slowing of biparietal diameter growth before 26 weeks menstrual age, had significantly reduced development quotient scores as assessed by the Ruth Griffiths scales. In our study there was a good correlation between the small fetus with a normal H/A circumference ratio and a low profile growth retardation curve, and this finding should raise the suspicion that the somatic deficit may be permanent. Elevated H/A circumference ratios, on the other hand, were associated with the late flattening growth retardation pattern or a normal serial cephalometry curve and in these cases Fancourt
174
CAMPBELL AND THOMS
et aZ(1976) showed normal catch-up growth and unimpaired intellectual development. REFERENCES Billington, W. D. (1975): European Journal of Obstetrics Gynecology and Reproductive Biology, 5, 147. Bonnar, J., Redman, C. W. G., and Sheppard, B. L. (1 975) : European Journal of Obstetrics Gynecology and Reproductive Biology, 5, 123. Bray, P. F., Shields, W. W., Wolcott, P. J., and Medsen, J. A. (1969): Journal ofPediatrics, 75, 303. Brettes, J. P., Renaud, R., and Gandar, R. (1976): American Journal of Obstetrics and Gynecology, 124, 2. Campbell, S. (1968): Journalof’ObstetricsandGynaecology of the British Commonwealth, 75, 568. Campbell, S. (1973): Lancet, 2, 1145. Campbell, S. (1974~):Clinics in Obstetrics and Gynaecology, 1,41. Campbell, S. (19746): Clinics in Perinatology, 1, 507. Campbell, S. (1976): Fetal Physiology and Medicine. Edited by R. W. Beard and P. W. Nathanielz. W. B. Saunders Co Ltd, London, pp 285,292.
Campbell, S., and Newman, G. B. (1971): Journal of Obstetrics and Gynaecology of the British Commonwealth, 78, 513. Campbell, S., and Wilkin, D. (1975): British Journal of Obstetrics and Gynaecology, 82, 689. Dobbing, J. (1970): American Journal of Diseases in Childhood, 120, 41 1 . Fancourt, R., Harvey, D. R., Norman, A. P., and Campbell, S. (1976): British Medical Journal, 1, 1435. Gruenwald, P. (1966) : American Journal of Obstetrics and Gynecology, 94, 1 1 12. Levi, S., and Erbsman, F. (1975): American Journal of Obstetrics and Gynecology, 121, 262. Robinson, H. P. (1973): British Medical Journal, 4, 28. Rosso, P., and Winick, M. (1974): Journal of Perinatal Medicine, 2, 147. Thompson, A. M., Billewicz, W. Z . , and Hytten, F. E. (1968): Journal of Obstetrics and Gynaecology of the British Commonwealth, 75, 903. Varma, T. R., and Curzen, P. (1973): Journal of Obstetrics and Gynaecology of the British Commonwealth, 80, 1086. Whitelaw, A. G. L. (1976): British MedicalJournal, 1,985. Willocks, J., Donald, I., Duggan, T. C., and Day, N. (1964): Journal of Obstetrics and Gynaecology of the British Commonwealth, 71, 1 1 .
ULTRASOU ND MEASUREMENT OF THE FETAL HEAD TO ABDOMEN CIRCUMFERENCE RATIO IN THE ASSESSMENT OF GROWTH RETARDATION BY
S . CAMPBELL, Professor * AND
ALISON THOMS, Research Assistant * Institute of Obstetrics and Gynaecology, Queen Charlotte’s Hospital, London, W6
Summary Fetal head and abdomen circumference measurements have been made using pulsed ultrasound techniques. The mean head to abdomen (H/A) circumference ratio with 95 per cent confidence limits was determined in 568 normal pregnancies from 17 to 41 weeks menstrual age. The mean ratio was 1-18 at 17 weeks, but decreased slow1,y until 29 weeks when the ratio was 1 .I1 ; thereafter there was a sharp fall in the mean ratio to I .01 at 36 weeks and 0.96 at 40 weeks. The H/A circumference ratio was also determined in 3 1 small-for-dates fetuses within one week of delivery. In all cases the fetal weight as predicted from the fetal abdomen circumference measurement was below the 5th centile weight for gestation. The ratio was above the 95th centile limit in 22 (71 per cent) of these fetuses and evidence is presented to suggest that the H/A circumference ratio can be used to distinguish between symmetrical and asymmetrical growth retardation.
to rationalize some of the more recent therapeutic regimens designed to improve fetal growth (Bonnar et al, 1975; Varma and Curzen, 1973; Brettes et al, 1976) and would permit a more realistic appraisal of the short-term and longterm risks of the small-for-dates fetus, thus helping to decide the optimal time for delivery. From serial ultrasonic cephalometry studies, Campbell (1974~) described the low profile and the late flattening growth retardation patterns of the fetal biparietal diameter which would appear to conform to the symmetrical and asymmetrical growth retardation categories. It is likely, however, that more information would be gained about the type of growth retardation if the relationship between fetal head and body measurements could be studied. Asymmetrical growth retardation is manifested by a high fetal brain to liver ratio (Gruenwald, 1966) and while these organs cannot be directly measured, ultrasound circumference measure-
IT is now recognized that small-for-dates babies are not a homogenous population and that at least two morphological groups can be distinguished: one in whiich there is symmetrical reduction in the size of all organs (symmetrical growth retardation) and the second in which the baby has a long wasted body and a relatively large brain which has been preferentially protected from the full effects of the growth retarding stimulus (asymmetrical growth retardation). The aet iological mechanisms, perinatal risks and long term prognosis appear to differ between the two groups (Campbell, 1974a) and it would tie of great value if a reliable antenatal means of distinguishing between these types of growth retardation could be found. For example, it would help
* Present address : Department of Obstetrics and Gynaecology. King’s College Hospital, Denmark Hill, London SE5 165
166
CAMPBELL AND THOMS
ment of the fetal skull and the fetal abdomen at the level of the liver can be performed and would appear to be a logical alternative. N o information on the fetal head to abdomen (H/A) circumference ratio is at present available and the following study was designed to establish whether the H/A ratio was of clinical value in distinguishing between symmetrical and asymmetrical growth retardation.
METHODS Circumference measurements of the fetal head and abdomen were made from B-scan echograms obtained with the Diasonograph 4102 (Nuclear Enterprises Ltd, Edinburgh) using a frequency of 2 - 5 mHz and a velocity setting of 1540 metres per second. Measurements of the biparietal diameter (BPD) were also made by the combined A and B-scan method of Campbell (1968) using a sound velocity of 1600 metres per second (Willocks et al, 1964). Head circumference The method of head circumference measurement has not previously been described. The aim of the technique was to obtain a horizontal section of the fetal head which included both
the biparietal diameter (coronal plane) and the occipito-frontal diameter (saggital plane). The first part of the procedure involved making longitudinal scans of the fetal body to determine on which side the fetal spine was positioned; this manoeuvre was of importance later when determining the degree of flexion of the fetal head on the spine. The second part of the procedure was identical to that described for measurement of the fetal BPD (Campbell, 1968) in which longitudinal scans were made to determine the angle of inclination of the fetal head to the vertical axis. Subsequently, transverse scans were made with the transducer tilted to this angle so that a horizontal section of the fetal head was obtained; this was recognized by the appearance of the midline echo and the widest fetal head diameter at right angles to the midline echo was the biparietal diameter. Due to flexion of the fetal head, however, this transverse scan usually displayed the suboccipitobregmatic diameter in the sagittal plane. The final part of the procedure therefore involved rotating the scanning gantry (the direction of rotation being determined by the position of the fetal spine) until the head appeared as an ovoid and echoes from the third ventricle were
FIG.1 Echogram showing a horizontal section of a fetal head suitable for circumference measurement. The third ventricle is clearly shown in the midline one third of the distance from the synciput.
ULTRASONIC ASSESSMENT OF FETAL GROWTH RETARDATION
167
detected in the midline, one-third of the distance from the synciput (Fig. 1). This ensured that the occipito-frontal heald circumference section had been achieved.
known pregnancy complications and all babies, when delivered at term, were considered by paediatric assessment to be of appropriate neurological development.
Abdomen circumference The method of abdomen circumference measurement has been described previously (Campbell and Wilkin, 1975). The section was a transverse one through the liver at the level of the umbilical vein.
Fetal growth retardation There were 30 patients who were delivered of a total of 31 small-for-dates babies which were studied antenatally with H/A circumference ratio measurements. A small-for-dates baby was defined as being on or below the 5th centile limit according to the tables of Thompson et a1 (1968), corrections being made for maternal parity and the sex of the child. In these cases H/A circumference ratio studies were made for one or more of the following reasons: (i) a small-for-dates uterus as assessed by palpation, (ii) the occurrence of a complication which might prejudice placental function, and (iii) a fall below the 5th centile in the growth rate of the biparietal diameter measured serially (Campbell and Newman, 1971). In all cases, the fetal weight as predicted from the fetal abdomen circumference measurement (Campbell and Wilkin, 1975) was considered to be on or below the 5th centile weight for gestation (Thompson et al, 1968).
Measurement When the required sections of the head and abdomen were obtained, circumference measurements were made from a Polaroid photograph by means of a map measurer. As the HewlettPackard 141 B-oscilloscope incorporated in the Diasonograph could not contain life sized (5/5) images in the third1 trimester, 3/5 and 4/5 size images were measured, appropriate corrections being made. The usual reduction in image size of 0.93 to 1, which is a feature of this oscilloscope, was compensated for by altering the focal distance of the camera.
PATIE~NTS The study was arranged in two parts. First, 568 measurements were obtained from 523 patients in normal pregnancy to establish the mean, 5th and 95th per cent confidence limits for the fetal H/A circumj'erence ratio in relation to menstrual age. Second, measurement of this ratio was then made in pregnancies in which fetal growth retardation was suspected. Normal pregnancy Our data consisted of 568 measurements obtained from 523 patients with normal singleton pregnancies. The pregnancy was considered to be normal when the patient was sure of the date of her last menstrual period and had regular 28-day cycles. In most cases, the fetal age had also been determined by ultrasonic measurement of the embryonic crown-rump length up to 12 weeks amenorrhoea (Robinson, 1973) or fetal biparietal diameter measurement between 12 and 20 weeks (Campbell, 19743) and these assessments were in agreement with the menstrual age as calculated from the last menstrual period. In addition, there were no
,
,
n = 568
30.8
z0 . 7I6
20
24
28
32
36
40
MENSTRUAL AGE (weeks)
FIG.2 Graphic representation of data in Table I showing mean H/A circumference ratios with 5th and 95th centile confidence limits from 17 to 42 weeks menstrual age.
168
CAMPBELL AND THOMS
RESULTS Normal data The mean, 5th and 95th centile limits for the H/A circumference ratio were calculated (Table I) and plotted (Fig. 2) in 2-weekly groups to smooth out fluctuations due to small numbers. The mean H/A circumference ratio was 1 .18 at 17 weeks and decreased slowly until 29 weeks when the mean ratio was 1 .11; thereafter there was a sharp fall in the mean
ratio to 1.01 at 36 weeks and 0.96 at 40 weeks. The range between the confidence limits remained fairly constant at 0.18. Fetal growth retardation All the 31 small-for-dates fetuses had H/A circumference ratio measurements made within seven days of delivery. In each case the ratio was above the mean of the normal range (Fig. 3 and Table 11).
TABLE I Mean fetal HIA circumference ratios with 5th and 95th centile limits related to menstrual age from 13 to 42 weeks; values have been combined in 2-weekly groupings to smooth out fluctuations due to small numbers (568 individual measurements)
___
_ _ ~ ~
~~
~
~
_
H/A circumference ratio Menstrual age (weeks)
13-14 15-16 17-18 19-20 21-22 23-24 25-26 27-28 29-30 31-32 33-34 35-36 37-38 39-40 4142
Number Of measurements
__
18 39 77 54 41 22 18 36 23 31 42 49 67 41 4
5th centile
Mean
95th centile
1.14 1.05 1.07 1.09 1.06 1.05 1.04 1.05
1.23 1.22 1.18 1.18 1.15 1.13 1.13 1.13
0-99 0.96 0.96 0.93 0.92 0.87 0.93
1.07 1.04 1.02 0.98 0.97 0.96
1.31 1.39 1.29 1.26 1.25 1.21 1.22 1.22 1.21 1.17 1.11 1.11 1.05 1.06 1 .oo
1
.lo
TABLE I1 Summary of clinical details of the 31 growth retarded fetuses related to whether the HIA circumference ratio was normal or abnormal
H/A circumference ratio above 95th centile (22infants, 21 pregnancies) No. Severe Hypertension Forceps delivery Emergency Caesarean section Elective Caesarean section Intrapartum fetal distress Apgar score of 5 or less Congenital fetal abnormality Perinatal mortality
5 5 4 4 9
lot
23 23 18 18 53*1 48
0
-
2
9
* Percentage of patients who were allowed to labour.
t Stillbirths excluded.
Per cent
H/A circumference ratio above mean but within normal range (9 infants, 9 pregnancies) No.
Per cent
1
11
0 1
1 1 3t 2 1
11 11 14*t 31 22
11
_
ULTRASONIC ASSESSMENT OF FETAL GROWTH RETARDATION
\
169
Most of the 22 babies were described as wasted on paediatric examination. Six babies had thicknesses of eight skin-folds measured according to the method of Whitelaw (1976); the sum of the skin-fold thicknesses was below
Y
L,
0.7
Ljlllllllllllll
lb
20
24
28
32
36
Patient C H
MENSTRUAL ACE ( w e e k s ) FIG.3 H/A circumference ratios of 31 small-for-dates fetuses related to the normal range and classified according to the type of BPD growth curve. All measurements taken = late retardation of within 7 days of delivery. BPD growth; o = low growth profile for serial BPD measurenients; 0- == low B:PD growth profile with terminal arrest of growth; X = normal BPD increase.
Ratios above the 95th centile In this group of 21 patients (22 fetuses) there was one intrauterine death and one neonatal death. Both babies were examined postmortem and death was thought to be due to growth retardation and asphyxia. Five patients had severe hypertension (a blood pressure of 160/100 mm IHg or more) and the clinical diagnosis in all five was pre-eclampsia. Four patients required emergency Caesarean section (see below), a lfurther five required delivery by forceps and one patient had an assisted breech delivery. Four patients had an elective Caesarean section to protect the fetus from the stress of labour. Seven patients (including the one with the twin pregnancy) had normal vaginal deliveries. Of the 17 patients allowed to labour, 9 developed fetal distress as defined by late deceleration patterns on a cardiotocograph or a fetal scalp blood pH of less than 7.25. Ten of the 21 liveborn babies in this group had an Apgar score of 5 or less at one minute.
- Twins
&rth weiqhtr 1.92 L I 73 kq a t 37 weeks Apqarr 8 in I ~ 7 i In
40
I6
20
24
28
32
36
40
MENSTRUAL AGE ( w e e k s )
201-1 MENSTRUAL AGE IN W t E K S
FIG.4 Elevated fetal H/A circumference ratios in a twin pregnancy. Both BPD growth curves were of the ‘late flattening’ type.
170
CAMPBELL AND THOMS
the lower limit of normal (Whitelaw, unpublished data) in all cases. Seventeen fetuses showed normal growth of the BPD until the end of the second trimester or later and then developed a sharp slowing of the growth rate, thus conforming to the late flattening growth pattern (see Fig. 4). One of them showed some catch-up growth of the biparietal diameter after hospital admission (Fig. 5). Two of the 22 fetuses had a normal growth pattern on serial BPD measurement and the results obtained in one of these patients is shown in Figure 6. One patient showed a low BPD growth profile pattern followed by catch-up growth. Three patients had a low BPD growth profile initially and then developed late cessation of growth; two of these mothers smoked 15 cigarettes per day throughout pregnancy (Fig. 7).
Ratios between the mean and 95th centile In the nine patients with H/A circumference ratios within this range, there was one emergency Caesarean section but no forceps deliveries for fetal distress in labour. There was one elective Caesarean section because of concern for the condition of the fetus. Of the 8 patients allowed to labour, one developed fetal distress and three of the nine infants born had an Apgar score of 5 or less at one minute. Two babies had congenital defects: one had cystic fibrosis and the other had congenital rubella; these babies were not described as wasted at paediatric examination. Two fetuses had skin-fold measurements performed (Whitelaw, 1976) with values about the mean for the gestational age. Six of the nine fetuses had a persistent low BPD growth rate from early in the second trimester conforming to the low growth profile (Fig. 8). One of these six fetuses died in utero at 29 weeks gestation. The cause of death was a concealed antepartum haemorrhage. The pathologist described the placenta as abnormally small and noted that growth retardation was present. Three of the nine fetuses showed late growth retardation; one was of a very minor degree while the mother of another developed sudden late onset
pre-eclampsia and was delivered by Caesarean section.
Patient B K Forceps del. for fetal distress Birth weiqht 2 . 2 kq a t 37 weeks
0 7
I6
20
24 28 32 36 MENSTRUAL ACE (weeks)
40
I lor
-111111111111 Ib
22 26 30 34 MENSTRUAL AGE IN WEEKS
18
38
FIG.5 Elevated fetal H/A ratios in a fetus whose BPD growth curve showed late retardation followed by ‘catch-up’ growth.
171
ULTRASONIC ASSESSMENT OF FETAL GROWTH RETARDATION
Patient S P
Patient C B
Birth weiqht 1.17 kq at
Birth weiqht I 76 kq
36 weeks
Apqar
0 7
-5
in 1
0.7
20 24 28 32 36 MENSTRUAL AGE (weeks)
I6
40
I6
20 24 28 32 36 MENSTRUAL AGE (weeks)
40
i l L1-UIh
18
22
26
30
34
38
MENSTRUAL. AGE IN WEEKS
FIG.6 Elevated fetal H/A circumference ratios associated with normal growth of the BPD.
15
18 22 2b 30 34 M?NSTRUAL ACE IN W F F K S
38
FIG.7 Elevated fetal H/A circumference ratios in a patient who was a heavy smoker. The BPD growth curve showed early retardation with terminal cessation.
172
CAMPBELL AND THOMS
Patient A D Normal vaqinal delivery B i r t h weiqht 2.62 kq at 4 0 weeks Apqar 9 in I
0.7 I
I6
I
20
I
I
I
I
I
I
1
1
24 28 32 36 MENSTRUAL AGE (weeks)
I
I
I
40
201-1 MENSTRUAL AGE IN WEEKS
FIG.8 Normal fetal H/A circumference ratio associated with symmetrical growth retardation. The BPD growth curve showed the ‘low profile’ pattern.
DISCUSSION While accepting that measurement of the BPD by the combined A and B-scan technique
is the most accurate and most reproducible fetal measurement which can be obtained antenatally (Campbell, 1973), we decided t o use the fetal head circumference measurement when determining the ratio between head and body. This was principally because the BPD is only a single dimension of the head and is frequently not truly representative of the total fetal head and (by inference) brain size. This is especially true in dolichocephaly when the narrow BPD gives a false impression of subnormal head size. We believe the head circumference measurement to be more representative of brain size and this is supported by our finding that normal weight dolichocephalic fetuses have head circumference measurements within the normal range. Furthermore, postnatal studies show a good correlation between head circumference measurements and brain development (Bray et al, 1969; Dobbing, 1970). Only Levi and Erbsman (1975) have previously discussed antenatal ultrasonic measurement of the head circumference but from their description of the technique and the illustrative echogram it is doubtful whether the occipitofrontal diameter was always obtained in the sagittal plane. For reproducible and accurate measurement of the head circumference, we believe that it is essential to make allowance for flexion of the fetal head and the position of the third ventricle in the midline makes a very efficient indicator as to whether or not the optimal section has been obtained. We have found the mean standard deviation of three independent head circumference measurements in the third trimester to be 1 .X mm (Campbell, 1976) which represents a 1 .O per cent error; therefore, although reproducibility of head circumference measurements is not as precise as that of biparietal cephalometry, we believe that the advantages of using the circumference measurement outweigh this small loss of precision. The advantages of measuring the fetal abdomen rather than the fetal thorax have been discussed by Campbell and Wilkin (1975) and Campbell (1976). In particular, we believe that by using the umbilical vein as a reference point the reproducibility of abdomen circumference measurements is improved ; the mean standard deviation of three
ULTRASONIC ASSESSMENT OF FETAL GROWTH RETARDATION
independent fetal abdomen circumference measurements in our laboratory is 2.95 mm (Campbell, 1976) which represents an acceptable 2 . 0 per cent error. In addition, sections at the level of the umbilical vein and fetal liver would seem to be particularly appropriate in studies on the small-for-dates fetus, for the fetal liver is the most affected organ in asymmetrical growth retardation. Our normal data graph shows a changing H/A circumference raticr from 17 to 41 weeks menstrual age. At 17 weeks the mean head circumference measurement was I 8 per cent larger than that of the abdomen, but between 17 and 30 weeks there was only a gradual fall in the ratio as bolh head and abdomen measurements were increasing rapidly. Between 30 and 36 weeks there was a sharp reduction in the mean ratio due to slowing of head growth which was not demonstrated by the abdomen circumference measurement so that during the 36th week the mean head and abdomen circumference measurements were equal. Although there was some deceleration of abdomen circumference growth from 36 weeks to term, slowing down of head circumference growth was even more pronounced so that in late pregnancy the normally nourished fetus usually had a head to abdomen circumference ratio of less than 1. It is interesting to note that although the range for both the head and abdomen circumference measurements widened towards term (Campbell and Thorns, unpublished data) the confidence limits for the head to abdomen circumference ratio remained constant. By dividing the 31 small-for-dates fetuses in our study into two groups, dependent on whether the H/A circumference ratio was above or below the 95 per cent confidence limit, we were able to detect clinical differences between the groups. The wasted baby is typically caused by utero-placental vascular insufficiency and the clinical features of the 22 infants (21 pregnancies) which showed elevated ratios suggested this mechanism; there was a high incidence of severe pre-eclampsia, fetal distress in labour, operative intervention and low Apgar scores (Table 11) while all six babies appropriately investigated had significantly reduced skin-fold thicknesses. The aetiological factors associated
173
with symmetrical growth retardation are more complex but it is known that chronic maternal undernutrition (Rosso and Winick, 1974) heavy maternal smoking, fetal infection and genetic or chromosome abnormalities (Campbell, 19740) can all produce the stunted baby; if none of these factors can be implicated then an immunological mechanism may be at work (Billington, 1975). Whatever the mechanism, utero-placental vascular insufficiency is less commonly associated with this type of growth retardation and the nine small-for-dates fetuses in our study who had normal H/A circumference ratios had a lower incidence of maternal hypertension, fetal distress and operative intervention (Table 11). Furthermore, two babies in this group who had skin-fold thickness studies made showed no evidence of wasting. The two infants with congenital defects had ratios below the 95th centile limit and the possibility of the stunted abnormal baby should always be considered when a small fetus with a normal H/A circumference ratio is diagnosed antenatally. The good correlation between the two H/A circumference ratio groups and the type of cephalometry growth retardation pattern may be of importance in the long-term assessment of the small-for-dates infant. Fancourt et al ( 1 976), in follow-up studies of small-for-dates babies who were measured antenatally by serial cephalometry and who were examined at a mean age of 4 years, have shown that the fetus with the longest period of growth restriction (the low growth profile group) retained the somatic deficit in height, weight and (to a lesser extent) head circumference. Furthermore, those infants, who as fetuses showed slowing of biparietal diameter growth before 26 weeks menstrual age, had significantly reduced development quotient scores as assessed by the Ruth Griffiths scales. In our study there was a good correlation between the small fetus with a normal H/A circumference ratio and a low profile growth retardation curve, and this finding should raise the suspicion that the somatic deficit may be permanent. Elevated H/A circumference ratios, on the other hand, were associated with the late flattening growth retardation pattern or a normal serial cephalometry curve and in these cases Fancourt
174
CAMPBELL AND THOMS
et aZ(1976) showed normal catch-up growth and unimpaired intellectual development. REFERENCES Billington, W. D. (1975): European Journal of Obstetrics Gynecology and Reproductive Biology, 5, 147. Bonnar, J., Redman, C. W. G., and Sheppard, B. L. (1 975) : European Journal of Obstetrics Gynecology and Reproductive Biology, 5, 123. Bray, P. F., Shields, W. W., Wolcott, P. J., and Medsen, J. A. (1969): Journal ofPediatrics, 75, 303. Brettes, J. P., Renaud, R., and Gandar, R. (1976): American Journal of Obstetrics and Gynecology, 124, 2. Campbell, S. (1968): Journalof’ObstetricsandGynaecology of the British Commonwealth, 75, 568. Campbell, S. (1973): Lancet, 2, 1145. Campbell, S. (1974~):Clinics in Obstetrics and Gynaecology, 1,41. Campbell, S. (19746): Clinics in Perinatology, 1, 507. Campbell, S. (1976): Fetal Physiology and Medicine. Edited by R. W. Beard and P. W. Nathanielz. W. B. Saunders Co Ltd, London, pp 285,292.
Campbell, S., and Newman, G. B. (1971): Journal of Obstetrics and Gynaecology of the British Commonwealth, 78, 513. Campbell, S., and Wilkin, D. (1975): British Journal of Obstetrics and Gynaecology, 82, 689. Dobbing, J. (1970): American Journal of Diseases in Childhood, 120, 41 1 . Fancourt, R., Harvey, D. R., Norman, A. P., and Campbell, S. (1976): British Medical Journal, 1, 1435. Gruenwald, P. (1966) : American Journal of Obstetrics and Gynecology, 94, 1 1 12. Levi, S., and Erbsman, F. (1975): American Journal of Obstetrics and Gynecology, 121, 262. Robinson, H. P. (1973): British Medical Journal, 4, 28. Rosso, P., and Winick, M. (1974): Journal of Perinatal Medicine, 2, 147. Thompson, A. M., Billewicz, W. Z . , and Hytten, F. E. (1968): Journal of Obstetrics and Gynaecology of the British Commonwealth, 75, 903. Varma, T. R., and Curzen, P. (1973): Journal of Obstetrics and Gynaecology of the British Commonwealth, 80, 1086. Whitelaw, A. G. L. (1976): British MedicalJournal, 1,985. Willocks, J., Donald, I., Duggan, T. C., and Day, N. (1964): Journal of Obstetrics and Gynaecology of the British Commonwealth, 71, 1 1 .
