The predictability of the small-for-gestational-age infant by real-time ultrasound-derived measurements combined with pulsed Doppler umbilical artery velocimetry

The predictability of the small-for-gestational-age infant by real-time ultrasound-derived measurements combined with pulsed Doppler umbilical artery velocimetry Emanuel Gaziano, MD, G. Eric Knox, MD, Gael P. Wager, MD, Richard P. Bendel, MD, DeborahJ. Boyce, RN, and jeanne Olson, RDMS

Minneapolis, Minnesota During a 15-month period 373 level II ultrasound examinations were performed in 256 high-risk patients. In addition, pulsed Doppler spectral recordings of blood flow in the fetal umbilical arteries were made. A systolic/diastolic ratio was then calculated for each fetus. Real-time ultrasound-derived estimated fetal weight with the use of biparietal diameter and abdominal circumference was also calculated. The estimated fetal weights were categorized by placing them in a percentile for gestational age according to published nomograms. Complete birth data and outcomes were obtained in all patients. Both the systolic/diastolic ratio and ultrasound-estimated fetal weight grouped by percentile ranking for gestational age were highly predictive (p = 0.001) of babies who were subsequently born small for gestational age. Seventy-nine percent of the infants small for gestational age had umbilical artery systolic/diastolic ratios ;;.4, whereas only 21% had normal systolic/diastolic ratios. Forty-three percent of the infants who were small for gestational age had ultrasound-estimated fetal weights ~10th percentile for the gestational age at which it was measured. Umbilical artery systolic/diastolic ratios, which reflect an increase in peripheral resistance in the placental circulation, showed a highly predictive and discriminatory index for the evaluation of the fetus suspected of having growth retardation. (AM J 0BSTET GYNECOL 1988;158:1431-9.)

Key words: Small for gestational age, real-time ultrasound, umbilical artery ratios, velocimetry

Recent studies suggest strong correlation between certain umbilical artery Doppler velocity waveforms and intrauterine growth retardation (IUGR).'- 5 Because increases in the systolic/diastolic ratio in velocimetry measurements reflect increased peripheral resistance in the umbilical circulation, this measurement may be an ideal physiologic correlate of impaired fetal growth." Furthermore, refinement of traditional real-time ultrasound techniques and standardization of curves for specific measurements have yielded increasingly more accurate predictions of fetal weight. 7 When fetal weight is estimated by ultrasound-derived measurements and correlated with the gestational age at the time of the examination, a sensitive and specific method of predicting IUGR is possible.• The purposes of the present study are to assess the impact of umbilical artery pulsed Doppler studies on the accuracy of IUGR prediction and to compare this technique with real-time ultrasound measurements. As

From the Perinatal Center, Abbott-Northwestern Hospital. Presented at the Fifty-fifth Annual Meeting of the Central Association of Obstetricians and Gynecologists, Tarpon Springs, Florida, October 15-17, 1987. Reprint requests: Emanuel Gaziano, MD, Abbott-Northwestern Hospital, 800 E. 28th St., Minneapolis, MN 55407.

new diagnostic modalities are introduced, it is of critical importance to examine whether their application confirms what is already available through less expensive or invasive means or whether such application adds additional accuracy to prediction of clinical events. Because many diverse factors can cause IUGR, specific biophysical measurements are not likely to be completely satisfactory. However, the addition of increasing numbers of technologies that measure different aspects of the fetal physiologic state are likely to lend increased understanding of the fetal dynamic process, leading to more sensible clinical strategies.

Method The Perinatal Center at Abbott-Northwestern Hospital, Minneapolis, Minnesota is a tertiary high-risk referral center and a member of a Food and Drug Administration-approved ATLI ADR Fetal Doppler Program (in conjunction with Advanced Technology Laboratories, Inc., Bothell, Wash.). The present study also received approval from the Institutional Human Research Committee of the hospital. From january 1, 1986 to March 13, 1987, 831 complete level II ultrasound examinations were performed. Examinations were conducted 373 times in 256 patients who, in addition to a level II scan, underwent spectral 1431

1432 Gaziano et al.

June 1988 Am J Obstet Gynecol

Table III. Primary level II examination findings

Table I. Level II examinations during pregnancy Frequency of level II examinations

1

2 3

4-8

Total

No.

%

181 49 18

71 19

8

7 3

256

100

Table II. Primary indications for level II ultrasound examinations Indications

Previous abnormal ultrasound Evaluate multiple gestation Suspect IUGR Abnormal maternal serum u-fetoprotein History of or suspect anomalies Other abnormalities Total

Findings

No.

%

Normal examination Ultrasound measures not consistent with gestational age Multiple gestation: discordance or other abnormality Triplet gestation: concordant growth Major abnormalities . Central nervous system Genitourinary Gastrointestinal Cardiac Other abnormalities

122 27

48

8

3

12

5

28 16 4

11

Total No.

11

9

6 2 3

30

__l!

256

100

%

48 34 29 27 22 96

10 36

256

100

19 13 11 11

recordings of blood flow in the fetal umbilical arteries. Level II scans require approximately 45 minutes of technician and physician time and include standard ultrasound measures and a complete anatomic survey of the fetus, placenta, amniotic fluid, and uterus. An Advanced Technology Laboratories Ultra-Mark 8 with either a 5 or 3 MHz sector scanner transducer was used for the level II examination. A 3 or 5 MHz pulsed Doppler transducer with a 1.5 mm 2 sample volume size was used to produce Doppler spectral recordings of flow in the umbilical artery. Recognition of the umbilical artery in the cord segment is by the characteristic velocity waveform shape and sound while there is simultaneous display of flow in the umbilical vein. For Doppler studies of the umbilical artery, fetal activity state was not a consideration other than for its influence on the selection of an optimum cord segment for study. Minimal power settings were used to produce satisfactory recordings. Spectral Doppler recordings were analyzed and a systolic/diastolic ratio was calculated for each fetus as previously described.' These data became part of the clinical record but were not used by clinicians for management decisions. In addition to the systolic/ diastolic ratios, the following information was obtained: maternal demographic information, gestational age at the time of the examination, primary and secondary level II findings, biparietal diameter (BPD), femur length, abdominal circumference, ultrasound-derived estimated fetal weight, percentile of ultrasound-derived estimated fetal weight for

Table IV. Maternal age and gestation at time of level II ultrasound examinations %

Maternal age (yr) 15-19 20-30 31-35 36-40 41-45 Total Gestational age (wk) 15-19 20-29 30-34 35-40 41-44 Total

13 163 61 18 1

5 64 24 7 0

256

100

34 99 87 35 _1

13 39 34 14 _Q

256

100

gestational age, ammotlc fluid volume, and interval from level II scan to delivery. The estimated fetal weight was derived with BPD or femur length and abdominal circumference according to nomograms developed by Shepard et al., 7 and the estimated fetal weight percentile for gestational age was determined from tables de'leloped by Jeanty et al. 9 Only data from the last ultrasound evaluation before delivery were used for the analysis. Complete birth information and outcomes were obtained on all patients and included the pediatrician's estimate of gestational age, the percentile birth weight for gestational age, route of delivery, newborn disposition to nursery type (intensive care unit, transitional, normal), Apgar scores at I and 5 minutes, and the designation of the newborn growth's status as appropriate for gestational age (AGA), small for gestational age (SGA), or large for gestational age (LGA). The latter was determined by standard nomograms developed by Lubchenco et al. 10 and

Predictability of SGA infant

Volume 158 Number 6, Part 1

Battaglia et al.l 2 Data were analyzed by a computer program that yielded means and, where appropriate, t tests, regression analysis, medians, standard deviations, X2 values, and probability values for each variable or group of variables tested.

1433

58

Results

A total of 373 examinations were performed in 256 patients. Overall, 90% of the patients had one or two examinations, and only 3% had four or more examinations (Table 1). Of the latter, most eXaminations were for follow-up of complex fetal structural defects. The interval from the last level II scan to delivery was """7 days in 26% of patients and """30 days in 56%. Data with respect to estimated fetal weight were calculated for gestational age at the time of the examination and placed in a percentile category for that gestational age, so that outcomes remote from the examination could be assessed. The relationship of a given abnormal systolic/ diastolic ratio noted at a point remote from delivery to outcome is also a critical consideration of the study. Table II lists the primary indications for the level II examinations. As expected, a previous abnormal level I scan was the most frequent indication ( 19% ), followed by ultrasound evaluation of multiple gestation (13%) and evaluation of suspected IUGR and an abnormal maternal serum a-fetoprotein level, accounting for 11% each. The remainder were a diverse group and included placental localization, estimation of gestational age, and prior history of fetal anomalies. The level II examination was abnormal in 52% of patients, with major fetal structural abnormalities (22%) and inconsistencies of gestational age measures (11 %) accounting for the largest groups (Table III). Other fetal abnormalities, discordant growth, and abnormal placentation largely accounted for the remainder. Almost one third of the patients were .,31 years old, and in 86% the last examination before delivery had been conducted at """34 weeks' gestational age (Table IV). Umbilical artery systolic/diastolic ratios were grouped as values of 2 to 3, 4, or .,5. Most authors consider values .,3 abnormaP Because a value of 4 is borderline, it was grouped separately, whereas values .,5 are considered abnormal irrespective of gestational age. Sixty percent of the patients showed systolic/diastolic values in the normal range, whereas 14% had systolic/diastolic values .,5. Of interest, both the systolic/diastolic ratio and the ultrasound-estimated fetal weight grouped by percentile ranking for gestational age were highly predictive (p = 0.001 and p = 0.001, respectively) for babies who were born SGA (birth weight """lOth percentile). A sys-

18th :;.tile or less

11th-49th l
58th-89th xt lle

98th-99th :;.t lle

Fig. 1. Ultrasound-estimated birth weight grouped by percentile estimated fetal weight for gestational age (solid bars) as compared with actual newborn birth weight (stippled bars).

tolic/ diastolic ratio .,4 was found in 79% of SGA babies. Forty-six percent of the SGA infants had markedly elevated systolic/diastolic ratios .,5. Normal systolic/diastolic ratios were noted in only 21% of the SGA babies. When the ultrasound-estimated fetal weight indicated a value """lOth percentile for gestational age, 43% of SGA were predicted. When the fetus was predicted by ultrasound-estimated fetal weight to be """lOth percentile for gestational age, only 2% of infants were AGA and 0% were LGA (Fig. 1). However, ultrasoundestimated fetal weight tended to underpredict the LGA infant. Although the actual percentage of LGA babies was 28%, ultrasound-estimated fetal weight (percentile for gestational age) predicted only 9%. Fig. 2 shows the relationship between systolic/ diastolic ratios and the SGA infant, and Table V lists the data for the above two predictor variables. Of the other ultrasound measures, systolic/ diastolic ratios correlated most highly and significantly with abdominal circumference (p = 0.00 1) and estimated fetal weight not grouped by percentile rank for gestational age (p = 0.001; Table VI). Significant correlations were also found for BPD, femur length, and ultrasoundestimated fetal weight. The Apgar scores at I and 5 minutes, nursery disposition, and route of delivery showed nonsignificant relationships with the systolic/ diastolic ratio. For newborn growth status, only the systolic/diastolic ratio and ultrasound-estimated fetal weight (percentile for gestational age) showed significant correlations. Abdominal circumference approached significance and BPD and femur length showed no relationships. SGA status correlated highly with Apgar scores at I and 5 minutes (p = 0.01), whereas route of delivery and nursery disposition were not correlated. Intergroup correlations were computed for SGA,

1434 Gaziano et al.

June 1988 Am J Obstet Gynecol

s:D Ratio 5-9

46

S:D Ratio 4

S:D Ratio 2-3

18

8

28

38

48

58

r.a

SGA INFANTS

Fig. 2. Systolic/ diastolic (S: D) ratios grouped by value and compared with the actual percent of SGA infants.

t) :

11:1:

s

I

.

Fig. 4. Abnormal systolic (S) to diastolic (D) ratio of 4.2: 1.

Fig. 3. Normal umbilical artery velocity waveform. The systolic (S) to diastolic (D) ratio is 2.8: I (normal).

AGA, LGA, and the total group. The systolic/diastolic ratio showed a strong correlation with the SGA group (r = - 0.31) but not with the other groups.

Comment Previous reports show that Doppler velocity recordings of the umbilical arteries provide information on the functional status of the villous circulation by a relatively rapid and noninvasive technique.'· 6 •12 • 13 Figs. 3 and 4 show our typical normal and abnormal waveforms with pulsed Doppler and indicate the lack of inflexion points noted in and characteristic of umbilical artery waveform tracings. Others have indicated that the percentage of error in analyzing such waveforms is low and acceptable, particularly when five waveforms are averaged. 14

Giles et al. 13 have demonstrated that abnormal flow velocity waveforms are associated with small artery obliteration in the placental tertiary villi. The latter are thought to represent the resistance arteries of the umbilical placental circulation. Because ofthis strong physiologic correlation with functional impairment leading to retarded fetal growth, considerable interest is directed toward clinical application. To date, few studies exist to compare Doppler umbilical artery ratio with other ultrasound-derived measures for IUGR prediction. Most research indicates that abnormal waveforms are characterized by differentials between the lower diastolic velocities compared with the systolic. Lowered diastolic velocities and occasional reversal of flow result in high systolic/ diastolic ratios, which are characteristic of increased placental resistance. These changes in the placental circulation are thought to precede the development of clinical growth retardation as well as the attendant changes in the state

Predictability of SGA infant

Volume 158 Number 6, Part 1

1435

Table V. Comparison of actual newborn birth weight designation to that predicted by umbilical artery systolic/ diastolic ratio and ultrasound-estimated fetal weight grouped by percentile ranking for gestational age Umbilical artery systolic I diastolic ratio Actual newborn growth status

SGA (n = 24) AGA (n = 151) LGA (n =55)

2-3

5 (21%) 96 (64%) 39 (71%)

I

4

8 (33%) 3,4 (22%) 10 (18%) X = 29.70 p = 0.008 4 d.f.

I

Ultrasound-determined estimated fetal weight

5-9

:s;1 Oth percentile

11 (46%) 21 (14%) 6 (11%)

10 (43%) 3 (2%) 0 (0%)

I

Between 11th and 89th percentile

11 (48%) 126 (89%) 41 (86%) X= 78.17 p = 0.001 4 d.f.

Table VI. Relationship between umbilical artery ratios and newborn growth status (SGA AGA = 151, LGA = 56) with real-time ultrasound and delivery variables Umbilical artery ratios (n = 253) Variable

Umbilical artery ratio Ultrasound-EFW (percentile for gestational age) Gestational wk at level II scan BPD Femur length Abdominal circumference Ultrasound-EFW Apgar score at 1 min Apgar score at 5 min Nursery type Delivery type

I

p

0.001 0.001 0.001 0.001 0.001 NS NS NS NS

?.90th percentile

I

2 (9%) 12 (9%) 7 (9%)

=

24,

Newborn growth status

r

p

-0.25 0.92 0.92 0.87 0.86 0.16 0.12 -0.08 0.10

0.001 0.001 NS NS NS NS NS 0.01 0.01 NS NS

I

r

-0.25 0.44 -0.08 0.04 0.09 0.16 0.16 0.21 0.18 -0.15

-0.o7

EFW, Estimated fetal weight.

of fetal oxygenation, which might be reflected in fetal heart rate data.'· 15 Initial clinical studies by Trudinger et a!. 16 indicate that 74% of SGA infants in their series showed abnormal umbilical artery ratios. These investigators also noted a higher correlation when umbilical artery Doppler studies were used for fetal surveillance when compared with traditional methods of fetal heart rate scoring (sensitivity of umbilical artery ratio of 60% versus 17% to 26% for fetal heart rate scores). 15 Fleisher et al. 2 grouped percentile birth weight into four categories (<25% being the lowest) and indicated a predictive value of a positive test (systolic/ diastolic ratio ~3) to be 49%, which increased to 66% in the presence of maternal hypertension. Erskine and Ritchie' 4 studied nine fetuses with severe IUGR over time. All nine had abnormal waveforms and four showed zero diastolic flow. Normal arterial waveforms indicate flow through diastole. A severe decrease in diastolic flow indicates markedly elevated placental

vascular resistance, and low or absent diastolic flow may suggest impending fetal death. As pregnancy progresses there is a small and gradual fall in umbilical artery resistance reflected as a decrease in the umbilical artery ratio. 14 This observation is well established and reported in most series. Of 62 abnormal waveforms reported in a series of high-risk patients, Trudinger et al. 3 demonstrated poor fetal outcome in 63%. Twenty-three of the infants with abnormal umbilical artery ratios had good fetal outcomes. Fleischer et al. 12 indicated mean ( ± SD) systolic/ diastolic ratios in the umbilical artery of 4.2 ± 1.1 in patients with preeclampsia and 3. 7 ± 1 in patients with chronic hypertension and preeclampsia. These values, in combination with abnormal uterine artery systolic/ diastolic ratios, were highly predictive for SGA infants. Recently that group further expanded their findings in hypertensive pregnancies, showing high correlations between systolic/ diastolic ratios in umbilical and uterine arteries with low birth weight and

1436 Gaziano et al.

adverse perinatal outcome. 17 Finally, Divon et al. 8 recently reported 111 patients at risk for IUGR. Umbilical artery ratios and ultrasound-estimated fetal weight showed positive predictive values for IUGR at 77.3% and. 76.3%, respectively. The present study confirms the high reliability of certain traditional ultrasound measures (estimated fetal weight), particularly when grouped by percentile for gestational age, in predicting IUGR. However, if one includes umbilical artery ratios ""4, 79% of the SGA babies had abnormal ratios compared with 43% of the SGA group whose estimated fetal weight was determined to be ~lOth percentile for gestational age. The umbilical artery ratio was associated with other strong ultrasound predictors of fetal weight such as abdominal circumference (p = 0.001) and ultrasoundestimated fetal weight (p = 0.00 1) and indicated a significant relationship with BPD and femur length. This observation reflects the relatively high predictor status for the systolic/diastolic ratio with respect to the birth weight (SGA) variables, and reconfirms the importance of abdominal circumference as a determinant for weight. In addition to . estimated fetal weight percentile grouped by gestational age, abdominal circumference was the only ultrasound variable that approached significance with respect to the SGA infant. Grouping of estimated fetal weight by gestational age using the Jeanty et al." nomogram to determine percentile rank is critical because this variable is highly predictive of the SGA infant (p = 0.001), whereas ultrasound estimated fetal weight alone was not predictive. In this study umbilical arte~ ratios alone were not predictors of neonatal outcome. However, the SGA group, 79% of which had abnormal systolic/ diastolic ratios, had extremely poor outcomes, with a disproportionate number being admitted to the neonatal intensive care unit as compared with the AGA and LGA groups. The examination-to-delivery interval was ""30 days in 44% of the patients. Of particular importance is the suggestion that the SGA infant may be predicted remote from delivery. The umbilical artery ratio did show a significant correlation with gestational age (p = 0.01), falling toward term as vascular resistance in the placental circulation further diminishes, as noted by previous investigators. This large clinical series poses questions that need further assessment. What differentiates these SGA infants with abnormal waveforms from those with normal findings? A more detailed look at the present data and future studies may indicate a subgroup of SGA infants who have no increased vascular resistance in the placental circulation. Certainly infants with specific inherent genetic defects may be SGA because of factors not

June 1988 Am J Obstet Gynecol

related to increased resistance to flow. The value of the systolic/ diastolic ratio may lie in separating these two important groups. Whereas traditional ultrasound measures may predict fetal weight relatively well, this observation is an after-the-fact event. At what point do abnormal waveforms suggest sufficient impedance of flow to be predictive of retarded growth, or are abnormal waveforms present only with established growth retardation? The answers to these questions rest with future clinical studies. REFERENCES l. Trudinger BJ, Giles WB, Cook CM, Connelly A, Thompson RS. Umbilical artery flow velocity waveforms in high risk pregnancy. Randomized control trial. Lancet 1987; I: 188. 2. Fleischer A, Schulman H, Farmakides G, eta!. Umbilical artery velocity waveforms and intrauterine growth retardation. AM J 0BSTET GYNECOL 1985;151:502. 3. Trudinger BJ, Giles WM, Cook CM. Flow velocity waveforms in the maternal uteroplacental and fetal umbilical placental circulations. AM J 0BSTET GYNECOL 1985; 152: 155. 4. Reuwer PJHM, Bruinse HW, Stoutenbeek P, Haspels AA. Doppler assessment of the feto-placental circulation in normal and growth retarded fetuses. Eur J Obstet Gynecol Reprod Bioi 1984;18:199. 5. FitzGerald DE, Stuart B, DrummJE, Duignan NMY. The assessment of the fetoplacental circulation with continuous wave Doppler ultrasound. Ultrasound Med Bioi 1984;10:371. 6. Trudinger BJ, Giles WB, Cook CM, BombardieriJ, Collins L. Fetal umbilical artery flow velocity waveforms and placental resistance: clinical significance. AMJ 0BSTET GvNECOL 1985;92:23. 7. Shepard MJ, Richards VA, Berkowitz RL, Warsof SL, HobbinsJC. An evaluation of two equations for predicting fetal weight by ultrasound. AM J 0BSTET GYNECOL 1982;142:47. 8. Divon MBY, Guidetti DA, Braverman JJ. Intrauterine growth retardation-a prospective study of the diagnostic value of real time sonography combined with umbilical flow velocimetry. Presented at the seventh annual meeting of the Society of Perinatal Obstetricians. Buena Vista, Florida: Society of Perinatal Obstetricians, 1987. 9. Jeanty P, Cantraine F, Romero R, Cousaert E, Hobbins JC. A longitudinal study of fetal weight growth. J Ultrasound Med 1984;3:321. I 0. Lubchenco LO, Hansman C, Boyd E. Intrauterine growth in length and head circumference as estimated from live births at gestational ages from 26 to 42 weeks. Pediatrics 1966;37:403. ' II. Battaglia FC, Lubchenco LO. A practical classification of newborn infants by weight and gestational age. J Pediatr 1967;71:159. 12. Fleischer A, Schulman H, Farmakides G, Brucero L, et al. Uterine artery Doppler velocimetry in pregnant WO!llen with hypertension. AM j 0BSTET GYNECOL 1986; 154:806. i3. Giles WB, Trudinger BJ, Baird PJ. Fetal umbilical artery flow velocity waveforms and placental resistance: pathologiccorrelation.Br J Obstet Gynaecol1985;92:3l. 14. Erskine RLA, Ritchie JWK. Umbilical artery blood flow characteristics in normal and growth retarded fetuses. Br J Obstet Gynaecol 1985;92:605. 15. Trudinger BJ, Cook CM,Jones L, Giles WB. A comparison of fetal heart rate monitoring and umbilical artery waveforms in the recognition of fetal compromise. Br J Obstet Gynaecol 1986;93: 171.

Predictability of SGA infant

Volume 158 Number 6, Part l

16. Trudinger BJ, Giles WB, Cook CM, BombardieriJ, Collins L. Fetal umbilical artery flow velocity waveforms and placental resistance: clinical significance. Br J Obstet Gynaecol 1985;92:23. 17. Ducey J, Schulman H, Farmakides G, Fleischer A, eta!. A classification of hypertension in pregnancy based on Doppler velocimetry. AM J 0BSTET GYNECOL 1987;157: 680.

Editors' note: This manuscript was revised after these discussions were presented. Discussion DR. FEDERICO G. MARIONA, Detroit, Michigan. The ability to predict prenatally the presence of IUGR has become the nemesis of the contemporary obstetrician. This is clearly demonstrated by the presence of an average of one clinical or research paper published on the subject every month in every major obstetric or perinatal journal. The authors do not report the frequency of such an occurrence in their institution, but one may assume that part of the impetus for their study was created by this complication. In our hospital, which is an urban institution caring for an underserved obstetric population, the walk-in patients produce 16.2% of the growthretarded babies, the clinic population produces 11.8%, and the private patients account for 4.5%. Growth alterations are indeed a new source of litigation discovered by the legal brethren and promptly utilized to attempt to demonstrate substandard obstetric care. 1 • 2 The authors have selected fetal vascular velocimetry to attempt to demonstrate this technique's superior ability to predict the SGA neonate when added to traditional imaging parameters. Indeed, this is a monumental undertaking and they must be commended for this effort, which represents physiologic rather than simplistic anatomic findings. Schulman' has recently emphasized the use of this noninvasive and apparently safe technique to enhance the understanding of fetal and maternal disease. The authors have also added "quick" to the previous advantages even when a complete examination may take 45 minutes. Dr. Gaziano and his colleagues have utilized a continuous wave/pulsed Doppler method that provided them with a concomitant image as they insonated the selected vessel. These systems use higher power levels that the authors consider "minimal," although they do not describe them. These same systems have been used overseas, where the power limits set by the Food and Drug Administration are not imposed. Why is it necessary to select a group in which to perform fetal Doppler studies? We all see neonatal Doppler studies performed daily in nurseries by pediatric cardiologists in the very same preterm or SGA infants we supply from our obstetric population. Is in utero Doppler insonation any different? Two hundred fifty-six pregnant women were studied. Of these, 34 (13.3%) carried a multiple pregnancy, and we assume that the authors have performed a dif-

1437

ferent set of calculations in this population to account for the slower BPD growth of twins in the third trimester! Table III shows that only 122 patients (48%) had a "normal" level II ultrasound examination. Clinical/sonic disparity, central nervous system anomalies, and "other" anomalies occurred in one third of their patients. Should consideration be given to include only fetuses from normal pregnancies while attempting to characterize the ability of a technique to predict a complication, because IUGR is known to occur in abnormal fetuses?' The study was prospectively executed and appropriately blinded for velocimetry results. Can the authors separate the predictive ability of the wave pattern in their three distinct populations: (1) patients scanned <7 days from birth, (2) patients scanned <30 days from birth, and (3) patients scanned >21 days from birth? Older and more recent studies5 • 6 emphasize the presence of an abnormally elevated systolic/ diastolic ratio in IUGR fetuses, a fact that these authors have reconfirmed, albeit once again in a small series. Is there at this time enough evidence to embark in cost-benefit studies and to move forward, discarding multiple sonic anthropometry in the detection of fetal growth retardation? If so, at which fetal age do the authors recommend the Doppler measurements, how far apart will they be necessary, and what is their expected negative predictive value? The authors made an extremely critical remark in emphasizing the importance of detecting an abnormal systolic/ diastolic ratio remote from birth. What is their definition of remote? Once IUGR is predicted, the differential diagnosis between symmetric and asymmetric growth retardation appears mandatory. Rapid karyotype, determination of fetal IgM levels, and differential miniallumbilical vascular wave patterns, which will indeed demand improvement in imaging resolution, may be essential. These steps appear to me extremely important to enable the obstetrician/ perinatologist to appropriately counsel the prospective parents, seek neonatology consultation, determine appropriate follow-up techniques, and plan delivery setting and modality. Indeed, the problem of iUGR taxes the ability of the clinician to suspect it, investigate it, predict it, and treat it appropriately. In treating pregnant women, the art is to deliver the growth-retarded fetu~ in a timely manner, before trophoblastic deterioration and placental senescence cause fetal death. The science is not accurate to predict and later detect lUGR and assist in making obstetric decisions. I encourage the authors to expand their study and limit it to singleton, no-risk fetuses. In the meantime, I am hopeful that the technique of Doppler waveform interpretation will weather appropriate scientific scrutiny and cost-benefit analysis and demonstrate its clinical usefulness. REFERENCES 1. Schwartz SS, Tucker ND, eds. Intrauterine growth retardation. Wiley Law Publications, 1985:269-76.

1438 Gaziano et al.

2. Elias S, Annas GJ, eds. Reproductive genetics and the law. Chicago: Yearbook Medical, 1987:89. 3. Schulman H. The clinical implications of Doppler ultrasound analysis of the uterine and umbilical arteries. AM ] 0BSTET GYNECOL 1987;156:889-93. 4. Yarkoni S, Reece EA, Holford T, et a!. Estimated fetal weight in the evaluation of growth in twin gestations: A prospective longitudinal study. Obstet Gynecol 1987;69: 636. 5. Trudinger BJ, Giles WB, Cook CM. Flow velocity waveforms in the maternal uteroplacental and fetal umbilical circulations. AM J 0BSTET GYNECOL 1985; 152:155-63. 6. McCowan LM, Erskine LA, Ritchie K. Umbilical artery Doppler blood flow· studies in the preterm, small for gestational age fetus. AMJ 0BSTET GYNECOL 1987;156:655-9.

DR. PAUL G. ToMICH, Chicago, Illinois. Dr. Gaziano and his associates have performed a clinical study to assess the impact of umbilical artery Doppler flow on the accuracy of IUGR predictions and to compare that technique with ultrasound measurements. Two hundred fifty-six patients underwent 373 level II ultrasound examinations; simultaneously, estimated fetal weight arid percentile ranking for gestational age and a systolic/ diastolic ratio were· calculated. That information became part of the clinical record but was not used by clinicians for management decisions. Birth dates and neonatal outcome were analyzed and statistical analyses were retrospectively performed. The important results of this study are as follows: (1) Sixty percent of the study group had systolic/ diastolic values within the normal range, 26% had systolic/ diastolic ratios >3 but <5, and 14% had a systolic/diastolic ratio ;;.5; (2) both the systolic/ diastolic ratio and ultrasound-estimated fetal weight grouped by percentile ranking were highly predictive of infants who were SGA; (3) the systolic/ diastolic ratio correlated best with abdominal circumference and estimated fetal weight; and (4) SGA status correlated best with neonatal outcome as measured by Apgar scores. The measurement of blood flow in both the fetus and the uterus by Doppler ultrasound represents a new and important advance in obstetrics. Its precise role in the surveillance of complicated pregnancies is now being established by studies such as this. The obstetric conditions most stt~died are those complicated by hypertension and suspect IUGR. With regard to IUGR, the promising potential of Doppler analysis is twofold: (1) early detection before other diagnostic tests and (2) blood flow changes determined as precedents of cardiotocographic signs of fetal hypoxia and acidosis. Stuart Campbell has performed pulsed Doppler screening at 16 to 18 weeks' gestation to determine if complications could be predicted by this measurement.' Laurin et al." studied fetal blood flow in pregnancies complicated by IUGR and have shown it to be an effective tool in the surveillance of those pregnancies. Fleischer et al. 3 have also shown the systolic/diastolic ratio to be a useful tool in managing pregnancies at risk for IUGR. From this study it would appear that if an infant is AGA or LGA, performing such a test would not be

June 1988 Am J Obstet Gynecol

indicated and have no additional predictive value, in addition to being costly. When IUGR is suspected or proved, determining a systolic/ diastolic ratio would be warranted, and that information would be useful in later clinical management decisions. I have three questions for Dr. Gaziano: How was it th~t the systolic/ diastolic ratio was part of the medical records but not used in clinical decisions? Concerning those SGA infants who had normal systolic/ diastolic ratios, on what basis were they determined to be SGA and what was their neonatal outcome? Lastly, did any of the SGA infants have absence of end diastolic flow? REFERENCES I. Campbell S, Pearce JM, Hackett G, et a!. Qualitative as-

sessment of uteroplacental blood flow: Early screening test for high-risk pregnancies. Obstet Gynecol1986;68:649-53. 2. Laurin J, Lingman G, eta!. Fetal blood flow in pregnancies complicated by IUGR. Obstet Gynecol 1987;69:895-902. 3. Fleischer A, Schulman H, Farmakides G, eta!. Umbilical artery velocity waveforms and IUGR. AM] OBSTET GYNECOL 1985;151:502-5.

DR. DANIEL A. RIGHTMIRE, Springfield, Illinois. I noticed that a number of the fetuses had multiple anomalies or at least major anomalies. Fetuses with cardiac abnormalities may have abnormal systolic/ diastolic ratios based on upstream abnormalities of flow in addition to any downstream flow abnormalities. I would like to ask the authors if better discrimination of SGA infants might not be obtained by classifying systolic/ diastolic ratios according to age-specific percentiles, just as they did for estimated fetal weight, because the normal range of systolic/ diastolic ratios varies with gestational age. The sample volume used was 1.5 mm2 • Small sample volumes might result in falsely elevated systolic/ diastolic ratios if that sample volume does not intercept the small column of red blood cells with the greatest velocity. If so, the systolic/ diastolic ratio would be more sensitive and less specific; this is what I believe the authors have shown. With regard to power output, the outputs and percentages for the Ultra-Mark 8 varied from about 10% to 12%. This would be expected to be within the range of the acceptable Food and Drug AdminiStration "derated" values. When power outputs were measured on my Ultra-MarkS, assuming a 50% measurement error, I co'uld stay within the "derated" values by increasing to 31.6%. Finally, I would ask the authors if they would consider displaying their data as a graph of the systolic/ diastolic ratio as a function of the gestational age at which that ratio was measured, plotted against the background of normal values. They could then highlight SGA, AGA, and LGA infants with different symbols. This would display all of their data in a form readily understandable to the observer. DR. EDGAR L. MAKOWSKI, Denver, Colorado. I think we must be. rather precise and careful in our termi-

Volume 158 Number 6, Part 1

nology. What we have to do is eliminate the term flow, because we are not actually measuring flow. To measure flow, one must know the angle of inclination and the diameter of the vessel that is being measured at a given point. All that we are measuring here is velocity. I have one question to ask the authors: did they consistently measure the umbilical artery velocity at a given point from either the abdominal wall of the fetus or from the placenta? DR. GAZIANO (Closing). In terms of Dr. Mariona's question relating to sample selection, we were performing Doppler studies to determine power output for the equipment. The sample selection was simply based on patients who were referred to the perinatal unit for level II examinations and who were generally a higher risk population than one would ordinarily see. I agree with the suggestion that such studies should be done prospectively in a low-risk population. The recommended fetal age at which to do the examination should be at the fetal age (preferably beyond 20 to 24 weeks' gestation) at which most cases of IUGR are clinically suspected. Obviously with those pregnancies that involve a high risk of IUGR such as hypertensive pregnancies, the timing of testing basically becomes a clinical decision. We did not look at the data with respect to the importance of certain predictions. Such data with respect to the interval from the examination to delivery would be useful and interesting. In fact, our study raised as many issues as it answered; there are considerably more data to analyze in the present study population. Dr. Tomich's remark with respect to the small fetus with a normal Doppler study is an intriguing question. Some of the recent data from the Netherlands and others indicate that the systolic/ diastolic ratio may become elevated before there are other changes in, for example, fetal heart rate data. It would be interesting to follow patients prospectively in that regard. I have no information on small fetuses with normal Doppler data. Yes, we did have groups in the series with no or reversed diastolic flow, and in most instances outcomes were poor. As to why we did not give the results to the clinicians during the study, we entered the study with absolutely no bias for or against umbilical artery Doppler data.

Predictability of SGA infant 1439

Because we had not used the technique previously, we felt that we would simply record the data, complete the study, and give the clinicians all the other ultrasound information. That was why we did not share the ratio at that point. Presently, the systolic/ diastolic ratio is on all of the records and reports that go to the referring physicians. Dr. Rightmire made some excellent points. We did have 22% major anomalies in our study group. We have not looked at the group with cardiac anomalies to determine the ratios separately for that group. His comment about the relatively small sample volume size taken in the umbilical artery indicating potentially less sensitivity is well taken. We find that we are able to do most of the studies with extremely low power settings of ,;;; 10% and yet obtain good waveforms. I particularly appreciate Dr. Makowski's comment on terminology. He is absolutely correct: velocity waveform is what we are talking about. These measurements are independent of angle, and flow determination requires angle measurements. There have been other studies in which the Doppler measurement was taken at different points in the umbilical cord, that is, near insertion to the placenta and also near fetal insertion. These different sites have shown no significant difference in calculated ratios. Furthermore, no differences in systolic/ diastolic ratios in umbilical artery are demonstrated with uterine contractions or certain of the fetal activity states. We did not average our results. If we had five waveforms that were similar, we measured the systolic/diastolic ratio. Actually, the difference is not significant whether we averaged them or not. Our results were highly reproducible, as one would imagine from the tracings. Finally, another comment about the English language. We used the terms IUGR and SGA interchangeably. Most clinicians think, talk, and use the term "IUGR." We confined IUGR to use as a generic term when referring to the fetus, and the term "SGA" is applied to the infant after birth. The report reads with the terminology more appropriate and fitting for the English language, but we are grateful for those who make these types of observations.