Umbilical flow waveforms versus fetal biophysical profile in hypertensive pregnancies

Umbilical flow waveforms versus fetal biophysical profile in hypertensive pregnancies

European Journal of Obstetrics & Gynecology and Reproductive Elsevier EUROBS Biology, 33 (1989) 199-208 199 00829 Umbilical flow waveforms versus...

678KB Sizes 1 Downloads 87 Views

European Journal of Obstetrics & Gynecology and Reproductive Elsevier

EUROBS

Biology, 33 (1989) 199-208

199

00829

Umbilical flow waveforms versus fetal biophysical profile in hypertensive pregnancies Enrico Ferrazzi ‘, Maria Bellotti I, Chiara Vegni *, Antonio Barbera Stefania Della Peruta ‘, Bianca Ferro 2, Giovanni Agostoni ’ and Giorgio Pardi ’

*,

’ Department of Obstetrics and Gynecology, University of Milan, Ospedale San Paolo, and ’ Department of Neonatology, University of Milan, Clinica ‘L. Mangiagalli: Milan, Italy Accepted

for publication

15 February

1989

Summary The pulsatility index (PI) of the umbilical arteries was measured in 40 hypertensive pregnancies. Doppler-velocimetric data were kept unknown to the clinical staff. An abnormal PI was found in 79% of cases in which an abnormal fetal growth in utero had been diagnosed by ultrasonographic measurements. Serial PI findings showed worsening figures in most of the cases with an abnormal fetal growth, irrespective of the last absolute value. Amniotic fluid estimation and PI data were significantly correlated. PI values were markedly abnormal in fetuses with non-reactive heart-rate tracings. A high sensitivity and an optimal specificity were found for umbilical PI versus the diagnosis of fetal growth retardation made by the coexistence of different biophysical criteria. However, false normal results may occur. 62% of the newborns weighed below the 5th percentile. The sensitivity of abnormal PI values to detect these light fetuses resulted to be only 67%. However the prevalence of neonatal morbidity in fetuses with abnormal PI values was 74%, while morbidity occurred only in 14% of cases with normal PI values. In hypertensive pregnancies, this simple velocimetric parameter proved to correlate with abnormal biophysical monitoring and complicated neonatal outcomes. Intra-uterine

growth

retardation;

Hypertension;

Doppler:

Umbilical

artery;

Flow velocity

waveform

Introduction The diagnosis of fetal growth retardation and its differentiation from constitutionally small fetuses as well as the evaluation of a proper timing of delivery in

Correspondence: Dr. Enrico Ferrazzi, Clinica Ospedale San Paolo, via Di Rudini 8,‘20142

002%2243/89/$03.50

Ostetrica e Ginecologica, Milano, Italy.

0 1989 Elsevier Science Publishers

Universita’

B.V. (Biomedical

degli Studi di Milano,

Division)

200

high-risk cases rely on the information provided by fetal biophysical parameters. However, sonographic estimation of fetal weight [l] and biometric evaluation of fetal growth patterns are limited by their inaccuracy [2]. Similar methodological and clinical limitations are reported for sonographic estimation of amniotic fluid volume [3]. On the contrary, fetal movements and cardiovascular reactivity can provide daily reliable information on fetal well-being as it is assessed by the fetal biophysical profile [4]. Indeed, the traditional criteria to diagnose fetal distress by fetal heart-rate monitoring may well be inadequate for proper obstetrical management. In recent years, pilot studies have introduced Doppler analysis of arteria-flow waveforms in the diagnostic approach to fetal growth retardation [5-71. These studies have raised either great expectations or skeptical comments. The aim of this prospective study was to compare the traditional biophysical information with the simplest among velocimetric techniques, i.e., the pulsatility index on the umbilical artery. The study was carried out in a blind way, such as not to influence the obstetrical management. Patients and procedures

The Pulsatility Index (PI) was measured on the umbilical arteries of 40 patients affected by pregnancy-induced hypertension (PIH). The last pulsatility index was obtained within 2 days from delivery. Serial measurements with a span interval of at least 2 weeks were obtained in 22 fetuses. In 15 cases the first velocimetric study was performed before 28 weeks. The equipment used was a co-axial pulsed Doppler with a sample volume from 9 to 25 mm. and high pass filters at 100 Hz (A.T.L. MK 600 and A.T.L. Ultramark 4). The lowest energy-output setting as possible was used. Examinations were performed from 9.00 to 12.00 a.m. with the patient in the semirecumbent position. Doppler samples were obtained preferentially with the fetus in a quite state. The umbilical vein-flow profile was checked to exclude breathing movements which modify to a great extent arterial velocimetric characteristics. The length of each examination very seldom exceeded 30 seconds. The pulsatility index was measured according to the simplified Gosling formula [8]. Three consecutive waveforms were measured on hard copies by means of a computerized map measurer (see Appendix 1). Velocimetric results were kept unknown to the clinical staff. Therefore the obstetrical management was never influenced by this new biophysical test. Fourteen patients had mild PIH (blood pressure serial measurements higher than 140/90 mmHg), 17 patients had moderate PIH (blood pressure higher than 100/150 mmHg). In two of these patients hypertension occurred only after 34 weeks. Seven patients suffered from severe PIH (blood pressure higher than 160/110 mmHg), two patients had major renal disease with hypertension from the 2nd trimester. The stadard drug used was clonidine, a prevalent central alpha-adrenergic blocker. Nifedipine, a calcium-antagonist, was added in severe cases. The evolution of the maternal hypertensive disease was monitored and treated according to common surveillance ecriteria and was taken into account together with the fetal condition for proper obstetrical management. In three cases only,

201

delivery was decided upon maternal conditions only. An eclamptic crisis did not occur in this series. Fetal growth and well-being was monitored by means of serial sonographic measurements of head and abdominal circumference, semiqualitative determination of amniotic fluid volume, fetal body and breathing movements, and cardiovascular reactivity as determined by fetal heart-rate recordings. A diagnosis of abnormal fetal growth was made when serial abdominal circumference measurements showed a flattening from normal values to below the 10th percentile of our standards [9] or when a single scan showed that the head-to-abdomen circumference ratio fell above the 90th percentile of normal standards [lo]; this finding was associated with oligohydramnios. Oligohydramnios was defined by the presence of fetal crowding and fluid pockets smaller than 4-5 cm [ll]. A reduction from the normal amount of amniotic fluid was also recorded but relied on subjective criteria. The presence or absence of fetal breathing movements during sonographic examination was reported. A non-reactive fetal heart rate was defined by the absence of two consecutive accelerations (15 beats per minute of at least 15 seconds) within 20 min of a recording lasting up to 120 min and a frequency bandwidth below 10 beats per min. Details on these criteria are reported elsewhere [12]. However, non-reactive fetal heart-rate recordings were regarded as ‘too late’ signs of fetal jeopardy. The coexistence of different biophysical criteria was required to establish a diagnosis of fetal growth retardation: flexing or arrest of abdominal circumference growth, or asymmetrical head-to-abdominal circumference ratio and reduction of amniotic fluid volume and reduction of fetal activity or/and loss of reactivity of fetal heat rate. Timing and mode of delivery were recorded and compared to biophysical data and velocimetric information. The percentile of weight for sex and gestational age was defined according to Italian standards [13]. Neonatal morbidity and mortality were independently checked by neonatologists. The following adverse events which required neonatal therapy were recorded: respiratory distress syndrome; neonatal seizures; coagulopathy; metabolic anomalies such as acidemia, hypoglicemia, hypokaliemia, hyperbilirubinemia; early and late neonatal mortality and its causes.

Results Perinatal biophysical monitoring Pulsatility index values measured on the umbilical arteries in this series of hypertensive pregnancies were compared to the normal cross-sectional standards obtained in our laboratory on 110 normal pregnancies [14]. In Fig. 1, the last measured PI is plotted against the week of gestation at delivery. An abnormal fetal growth in utero was diagnosed by sonography for 25 fetuses. Abnormal PI values were observed in 19 of these fetuses. A concordance between the abnormal results of the two diagnostic approaches occurred in 79% of cases. In the other 6 fetuses with abnormal growth profile, the pulsatility index was within the normal ranges. In Table I the different biophysical parameters and the neonatal outcomes are reported

202

UMBILICAL

P I

4T--rrF

IN 40

HYPERTENSIVE

;4 WEEKS

;6

PATIENTS

;a

OF GESTATION

Fig. 1. The pulsatility index measured on the umbilical artery within 2 days from delivery is plotted against the normal standard (mean a! 2 SD). (See text for diagnostic criteria of abnormal fetal growth.)

in detail for these latter discordant cases. None of the 15 fetuses with normal growth at sonography showed abnormal PI values. Serial measurements of pulsatility index were obtained in 22 pregnancies of the whole series. Increasing PI values were found in 10 of 11 fetuses with a sonographic diagnosis of abnormal growth. On the contrary, the umbilical pulsatility index decreased in 8 out of 11 fetuses normal at sonography. In Fig, 2 the difference between the last minus the first umbilical PI is correlated with the severity of growth retardation observed at birth. The regression is highly significant (p -C0.003; r =

TABLE

I

Sonographic growth =

Head/ abdomen ratio

Amniotic fluid

Fetal heart rate

Last PI

Week of delivery

Newborn weight b (g)

Neonatal outcome ’

Flat

asymmetrical asymmetrical symmetrical asymmetrical

reduced oligohydr. reduced oligohydr.

reactive reactive reactive reactive

1.1 1.2 1.0 1.0

32 33 34 35

1140 (-46) 1220 (-48) 1650(-35) 1540(-43)

RDS Coagulo-

symmetrical asymmetrical

reduced oligohydr.

reactive reactive

1.0 1.0

36 38

1350(-56) 2100(-39)

-

Low P. Flat

Hat

a For serial examinations only. Flat = late flattening of fetal growth. Low P = fetal growth constantly below the 5th percentile. b Number in brackets shows the percentage of weight. Below the 50th percentile for gestational age. c Neonatal morbidity and mortality. No neonatal deaths in this group.

pathy

203

7 a

SERIAL

.6

P I vs GROWTH

RETARDATION

J ei

. :

a

z g

m i

-.41 -80

-6r,

PERCENTAGE

-25

-40

Of

WEIGHT

BELOW

THE

0

50TH

PERCENTILE

Fig. 2. The values of the difference of the last minus the first PI measurement, obtained at least 4 weeks earlier, is correlated with the percentage of weight below the 50th percentile at birth. The difference is corrected for the weekly reduction of PI vaIues observed in cross-sectional standards. Growth retardation is expressed as a percentage of the observed weight below the mean expected weight to provide a gestational-age independent parameter. The linear regression between the two parameters is highly significant (p < 0.003; r = 0.35).

0.35). In fact, fetuses showing increasing PI values had consistently lower weights at birth. The correlation between PI values and sonographic estimation of head to abdominal circumference ratio is shown in Fig. 3. The data are presented for two

UMBILICAL

P I AND

ASYMMETRICAL

GROWTH

0 .9

I

HEAD

CIRCUMFERENCE

1.1

1.2

/

ABDOMINAL

I.3

1.4

CIRCUMFERENCE

Fig. 3. Correlation of PI measurements and head-to-abdomen circumference ratio. Full squares (m) represent fetuses delivered within 34 weeks, empty squares (0) represent fetuses delivered from 35 weeks to term. The linear regression between the two parameters is highly significant for the whole population (p -C0.0001; r = 0.42) and for the two groups (p -C0.05; r = 0.33) (p < 0.0065; r = 0.39).

204

5-

PULSATILITY

INDEX AND AMNIOTIC FLUID I I

8

NORMAL

REDUCED

- OLICOHYDRAMNIOS

Fig. 4. F’ulsatility index measurements are shown for cases with normal amniotic fluid, and for cases with reduced amniotic fluid or oligohydrammios, dots. The Mann-Whitney U test is significant (p -C0.01).

different time periods, from 28 to 33 weeks of gestation and from 34 to term. A significant regression line was observed for the two subgroups. As expected, the regression between PIs and this index of asymmetrical growth retardation is steeper in earlier gestation. In Fig. 4, the pulsatility index of the umbilical artery is compared to the sonographic estimation of amniotic fluid volume. Abnormal PIs were significantly associated with the sonographic estimation of a reduced amniotic fluid volume or oligohydramnios (Mann-Whitney U test, p < 0.01). The highest values were consistently associated with marked oligohydramnios; however, a wide overlap was found between normal and abnormal conditions. Antepartum fetal heart rate tracings were obtained in all cases. According to our clinical protocols, in only 8 cases delivery was delayed until a non-reactive fetal heart-rate recording occurred. These fetuses were delivered by Cesarean section. Table II presents time of delivery, newborn weight, neonatal outcome and the pulsatility index of these 8 cases. Perinatal outcomes Sixteen (40%) fetuses were delivered before 34 weeks of gestation and thirteen (33%) before 37. All these cases were delivered by Cesarean section. Only three abdominal deliveries were performed for maternal indication. Vaginal deliveries were allowed only after 37 weeks of gestation in seven patients (17%). In this series 28 (70%) of the newborns weighed below the 5th percentile. The sensitivity of abnormal PI values to detect a neonatal weight below the 5th

Amniotic fluid

oligohydr. oligohydr. normal normal oligohydr. normal normal oligohydr.

Head/abdomen ratio

asymmetrical symmetrical asymmetrical symmetrical asymmetrical symmetrical symmetrical symmetrical

Flat Flat Flat Law P. Flat Flat Low P. Flat

a For serial examinations only. Flat = late flattening of fetal growth. Low P. = fetal growth constantly below the 5th percentile. b Number in brackets shows the percentage of weight. Below the 50th percentile for gestation& age. ’ Neonatal morbidity and mortality. Hypogl. = hypoglicemia. Coag. = coagulopathy. E. Hem. = endoventricular hemorrhage.

Sonoeraphic

growth a’

TABLE II Last PI

2.4 2.7 1.6 2.0 2.9 1.4 1.6 2.2

Fetal heart rate

non-reactive non-reactive non-reactive non-reactive non-reactive non-reactive non-reactive non-reactive

28 29 30 30 31 31 34 35

Week of delivery 720 (-28) 560 ( - 60) 890 ( - 47) 1150 (-31) 710 ( - 64) 1200 (- 38) 1800-34) 1700 (-40)

(8)

Newborn weight b

RDS, E. Hem., Sepsys, death Hypogl., Coag. Hypogl.

Hypogl., Coag., E. Hem., Death RDS, Coag., Death Primary resuscitation _

Neonatal outcome

206

percentile resulted in only 65%. Sensitivity was as high as 88% when worsening PI findings were taken into consideration independently from their absolute values. The prevalence of neonatal morbidity (10 cases) and mortality (4 cases) in the 19 fetuses with abnormal PI values was 74% (14 cases). Morbidity occurred only in 14% (3 cases) of fetuses with normal PI values. No neonatal death occurred in this group. A multiple regression was performed to test the correlation between mortality, morbidity or normal outcomes and gestational age (p < 0.62) newborn weight (p < 0.08) and the last PI value (p < 0.01). Early neonatal death occurred in 4 cases out of 7 with PI values higher than the 4th stardard deviation, approximately above 1.8, 2. Discussion The introduction of a new test in clinical management of pregnancies with growth retardation needs to be validated with other tests which have gained a well established role and with the gold standard for each of these tests, i.e., perinatal morbidity and mortality. In our opinion the two comparisons provide an indirect proof of the diagnostic value of the new information we are now able to obtain. A methodological condition required by this approach is that the new information we are obtaining must not influence the clinical management under any circumstance, i.e., must be unknown to the clinical staff. This was in fact the first step in the design of our study. We adopted the pulsatility index as a parameter of placental impedance to umbilical flow, for it takes into consideration not only the maximum and minimum velocity but also the mean velocity and the duration of the pulse. A single abnormal velocimetric test obtainable in a few minutes achieved a concordance of 79% with the diagnosis of abnormal fetal growth which is usually made by means of serial sonographic measurements of the fetus and or by the assessment of different biophysical criteria. This role of velocimetry is further stressed by the finding of a significant correlation (p < 0.001) between asymmetrical fetal growth and the pulsatility index on the umbilical arteries. Moreover, a significant correlation (p < 0.003) was found between growth retardation observed at birth and serial PI measurements, as is shown in Fig. 2. These latter results obtained on serial measurements need to be confirmed on a larger number of cases, but hint at a possible serial use of umbilical velocimetry at 18-22 weeks and later in the third trimester together with routine sonographic examinations in order to improve its sensitivity in detecting fetal growth retardation due to placental vascular damage. The comparison of PI with sonographic estimation of amniotic fluid suggests that these parameters are also significantly correlated. However, amniotic fluid estimation can be considered a gross marker of fetal kidney and placental function and anomalies in velocimetric data occur earlier than sonograpbic evidence of amniotic fluid reduction. We were able to compare PI measurements and non-reactive fetal heart rate only in 8 cases. In fact, our protocol tends to deliver growth-retarded babies before they

show stable non-reactive heart rates. All these fetuses had abnormal PI values ranging from 1.6 to 2.7. These observations are in agreement with other findings recently reported [6]. Different gestational ages, but more likely different compensatory mechanisms, play a major role in this wide range of placental impedance at the time fetal cardiovascular alarm condition is detected. Therefore a fixed abnormal PI value above which fetal reactivity is lost does not exist. However, a risk field can be considered, and this is in accordance with biochemical data obtained from the growth-retarded fetus in utero [16]. The diagnostic role of umbilical velocimetry is further emphasized by the correlation of this index with the neonatal outcome. The significantly higher prevalence of neonatal morbidity (53% vs. 14%) and mortality (21% vs. 0%) in newborns with abnormal PI then in newborns with normal PI values alerts the clinician that above that value the fetus can be damaged in utero by insufficient metabolic support. In the clinical management timing of delivery must therefore be taken into consideration among the different management solutions and balanced with the problems of prematurity and the neonatologist point of view. This approach to the clinical use of umbilical PI is supported also by the absence of false abnormal cases observed in this series. However, its combined use with other biophysical tests is still to be recommended. In fact, in our experience there were six fetuses with normal umbilical PI but with abnormal biophysical data, which allowed us to diagnose these fetuses as growth retarded (Table I). Three of these fetuses showed complicated neonatal courses. A solution to these findings can be found in the future by means of a more complete velocimetric investigation of the fetus and the placenta. In some reported series [17] and in our recent experience, the cerebral circulation and aortic flow characteristics seem to provide a higher sensitivity of fetal distress. Moreover, abnormal velocimetric profiles of the major uterine arteries can detect abnormal placental function even with normal umbilical velocimetry [18]. Appendix 1 The mean velocity calculated using a map measurer is obtained dividing the area of the flow waveform by the length of the cycle. The mean velocity computed by most of the automatic analyzers of the waveform is the sum of the maximum velocity of each sample of the Doppler spectrum analysis divided by the number of samples performed for each cardiac cycle. The higher the difference between maximum and minimum velocities, the greater the difference of the PI values measured by the two methods. The correlation between the two methods is defined by the following coefficients: automatic

PI = - 0.34 + 1.49 * (geometric

PI) - O.OSO(geometric

PI)*2

References 1 Hadlock FP, Harrist RB, Sharmann RS, Deter RL, Park SK. Estimation of fetal weight with the use of head, body and femur measurements - a prospective study. Am J Obstet Gynecol 1985;151:333-337.

208 2 Nicolini U, Ferrazzi E, A Kustermann, G Pardi. Accuracy of an average ultrasonic laboratory in measurement of BPD, head and abdominal circumference. 3 Perinat Med 1986;14:101-107. 3 Halperin ME, Fong KW, Zalev AH, Goldsmith CH. Reliability of amiotic fluid volume estimation from ultrasonograms: intraobserver and interobserver variation before and afer the establishment of criteria. Am J Obstet Gynecol 1985;153:264-267. 4 Marming GA, Morrison I, Lange IR, Harman CR, Chamberlain PF. Fetal assessment based on fetal biophysical profile scoring: experience with 12620 referred high-risk pregnancies. Am J Obstet Gynecol 1985;15:343-350. 5 Fleischer A, Shulman H, Farmakides G, Bracero L, Blattner P, Randolph G. Umbilical artery velocity waveforms and intra-uterine growth retardation. Am J Obstet Gynecol 1985;151:502-505. 6 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;93:171-175. 7 Hacket GA, Campbell S, Gamsu H, Cohen Overbeck T, Pearce JMF. Doppler studies in growth-retarded fetus predict neonatal necrotizing enterocolitis, and hemorrhage and neonatal morbidity. Br Med J 1987;294:13-16. 8 Gosling RG, King DH. Continuous wave ultrasound as an alternative and complement to X-rays in vascular examinations. In Reneman RS, ed. Cardiovascular application of Ultrasound. Amsterdam: North-Holland Publishing Company, 19’74;266-282. 9 Todros T, Ferrazzi E, Groli C, Nicolini U, Parido L, Pavoni M, Zorxoli S, Zucca S. Fitting growth curves to head and abdomen measurements of the fetus: a multicentric studies. J Clin Ultrasound 1987;15:95-106. 10 Campbell S, Thorns A. Ultrasound measurement of the fetal head-to-abdomen circumference ratio in the assessment of growth retardation. Br J Obstet Gynaecol 1977;84:165-169. 11 Philipson EH, Sokol RI, William T. Oligohydramnios clinical association and predictive value for intra-uterine growth retardation. Am J Obstet Gynecol 1983; 147:271-276. 12 Ferrazzi E, Bozzetti P, Pardi G. The oxytocin stress test: a re-evaluation. Perinatal medicine, problems and controversies. Di Renzo GC, Hawkins DF, eds. New York: Raven Press, 1985;19-24. 13 Consiglio Nazionale delle Ricerche. Standard de1 peso de1 neonato italiano. Riv Ital Ped (I.J.P.) 1980;6:153-170. 14 Ferrazzi E, Bellotti M, Della Peruta S, Gementi P, Agostoni G, Bulfoni G, Pardi G. Velocimetria Doppler: diagnosi differenziale tra feto piccolo e dificit di accrescimento endouterino. Ann Ost Ginecol Med Perin 1988;2:9C-98. 15 Giles WB, Trudinger BJ, Baird PJ. Fetal umbilical artery flow velocity waveforms and placental resistance; pathological correlation. Br J Obstet Gynaecol 1985;92:31-38. 16 Ferrazzi E, Pardi G, Buscaglia M, Makowski EL, Battaglia FK. The correlation of biochemical monitoring versus umbilical flow velocity measurements of the human fetus. Am J Obstet Gynecol 1988;159:1081-1087. 17 Arduini D, Rirzo G, Mancuso S, Romanini C. Longitudinal assessment of blood flow waveforms in the healthy human fetus. Prenat Diagn 1987;7:234-238. 18 Schulmann H. The clinical implications of Doppler ultrasound analysis of the uterine and umbilical arteries. Am J Obstet Gynecol 1987;156:889-893.