Doppler Echocardiographic Studies of Diastolic Function in the Human Fetal Heart: Changes During Gestation

Transcription

1 JACC Vol. 8, No Doppler Echocardiographic Studies of Diastolic Function in the Human Fetal Heart: Changes During Gestation KATHRYN L. REED, MD,* DAVID J. SAHN, MD, FACC,t SARAH SCAGNELLI, RDMS,t CAROLINE F. ANDERSON, RDMS,* LEWIS SHENKER, MD* Tucson, Arizona and San Diego, California With the combined use of two-dimensional ultrasound and Doppler echocardiography, noninvasive examination of the human fetal heart and circulation has recently become possible. These techniques were employed to investigate diastolic atrioventricular valve ftow in the fetal heart in 120 fetuses studied between 17 and 42 weeksof gestation. Two-dimensional ultrasound was used to examine fetal and intrauterine anatomy, and estimates of gestational age were made based on biparietal diameters and femur lengths. Doppler echocardiography was performed with a 3.5 or 5 MHz Doppler sector scanner. Flow velocity patterns were obtained through the tricuspid and mitral valves during diastole. Peak ftow velocity during late diastole or atrial contraction (A) was compared with peak ftow velocity during early diastole (E) in four groups of fetuses: Group 1, 17 to 24 weeks of gestation; Group 2,25 to 30 weeks; Group 3, 31 to 36 weeks; and Group 4, 37 to 42 weeks. The ratio of A to E decreased significantlyas gestational age advanced, from 1.56 ± 0.06 (± SE) to 1.22 ± 0.03 across the tricuspid valve (p < 0.001) and from 1.55 ± 0.04 to 1.22 ± 0.06 across the mitral valve (p < 0.001). In tricuspid valve measurements, peak ftow velocity during early diastole increased from 26.3 ± 2.0 cm/s in Group 1 to 36.5 ± 1.7 cm/s in Group 4 (p < 0.001), whereas peak ftowvelocityduring atrial contraction did not change. Mitral valve peak ftow velocityduring early diastole did not change significantly, whereas peak ftow velocity during atrial contraction decreased from 45.8 ± 1.3 cm/s in Group 2 to 34.5 ± 2.7 cm/s in Group 4 (p < 0.001). Tricuspid ftow velocities during early diastole and late diastole were greater than mitral ftow velocities in 91 fetuses in which both valves were studied (p < ), thereby providing further evidence of right heart dominance in the human fetus. Diastolic ftow into the left and right ventricles in human fetuses is shifted into late diastole when compared with newborns or adults. Doppler echocardiography is a usefultechnique for the study of human fetal cardiovascular function. (J Am Coli Cardiol1986;8:391-5) The availability of two-dimensional, M-mode and Doppler echocardiography has made possible noninvasive examination of human fetal cardiac anatomy and circulation. These techniques have been used to identify and study normal fetal cardiac anatomy and function in middle and late gestation (1-4). More recently, Doppler flow velocity patterns have been examined in the normal human fetal heart (5). Fetal cardiac anatomy differs from that in the adult in several ways. In the fetus, the foramen ovale and the ductus From the *Department of Obstetrics and Gynecology, Arizona Health Sciences Center, Tucson, Arizona and tdivision of Pediatric Cardiology, University of California San Diego Medical Center, San Diego, California. This work was supported by a grant from the American Heart Association, Arizona Affiliate, Phoenix, Arizona. Manuscript received June 25, 1985; revised manuscript received February 25, 1986, accepted March 17, Address for reprints: Kathryn L. Reed, MD, Department of Obstetrics and Gynecology, Arizona Health Sciences Center, 150I North Campbell Avenue, Tucson, Arizona arteriosus are normally patent. In addition, the right ventricle is slightly larger than the left (1), and the right ventricle is more hemispherical in shape than in the adult, as shown in both animal (6) and human (7) studies. Also, in the fetus blood flows through the ventricles in parallel rather than in series, with minimal blood flow through the pulmonary arteries (8). Ventricular compliance in fetal lamb hearts has been shown to be decreased when compared with that of newborn or adult sheep hearts (9). Values for fetal blood pressure, heart rate, hemoglobin, hematocrit and arterial and venous partial pressure of oxygen and oxygen saturation all differ from values in adults (10). Most of our knowledge regarding fetal circulation comes from studies performed in animals because, to date, this type of investigation has required invasive procedures. The purpose of this investigation was to use Doppler echocardiographic methods to measure diastolic flow velocity patterns across the tricuspid and mitral valves in human fetuses during gestation by the American College of Cardiology /86/$3.50

2 392 REED ET AL. FETALDOPPLER FLOWS IN DIASTOLE lacc Vol. 8, No.2 Methods Study patients. Patients were selected from the obstetrics clinic at Arizona Health Sciences Center and from outside referrals. Permission was obtained from the Human Subjects Committee to perform this study. Informed consent was obtained from each participating mother before every examination. Examinations began with a brief history of the current pregnancy. The fetus and intrauterine contents were.then studied with a 256- or 280-element 3.5 or 5 MHz linear array ultrasound scanner. Fetal number, lie, biparietal diameter, abdominal diameters, femur length and general fetal anatomy were examined. Amniotic fluid volume and placentallocation were also noted. Estimates of gestational age were made based on biparietal diameters and femur lengths (11,12). A 3.5 or 5 MHz two-dimensional echocardiographic Doppler Honeywell-Electronics for Medicine sector scanner was then used to identify cardiac chambers and determine locations for Doppler single range gate velocity interrogation. Regardless of the transducer frequency for imaging with this system, Doppler interrogations were ~erformed with 3.5 MHz with a fast Fourier transformation spectral velocity output. Both the linear array and the sector Figure 1. Calculation of NE ratio across t~e tric.uspid valv~ in the human fetus. A == peakflow velocity during atnal contracnon: E == peak flow velocity in early diastole. scanner operate at power outputs of less than 100 mw/cm 2 spatial peak-temporal average in both imaging and Doppler modes by manufacturers' specifications. Doppler echocardiographic methods. Normal fetal cardiac anatomy was confirmed by obtaining four chamber, short-axis/great vessel, aortic arch and ductal views. All fetuses were also examined after birth and confirmed to be normal. For obtaining flow velocity patterns through the tricuspid valve and the mitral valve regions, the location of the pulsed Doppler gate (0.1 to 0.3 em in length) was established using the two-dimensional image both before and after the Doppler tracing was obtained. Using the two-dimensional image in the four chamber view as a localizer, the pulsed Doppler gate was placed just distal to the leaflets of the mitral or tricuspid valves, and the flow velocities were maximized by fine transducer rotation and angulation. The angle between the Doppler beam and blood flow direction was estimated by assuming that blood flow was perpendicular to the valve anulus. Once the angle cursor was set manualiy, angle corrections were performed automatically by the Honeywell scanner, which divided the velocities by the cosine of the angle of incidence. Peak flow velocities in early diastole (E) and peak flow velocities in late diastole or atrial contraction (A) were calculated using a graphics tablet, linked to a computer, by measuring the highest modal velocity (the darkest part of the gray scale spectral output) (Fig. I). An average of three beats was used to establish each value. Doppler studies were performed by one of three examiners (5.5., C.F.A. B CI'I ANGL (d.;r ) = un CI'I/. 6 el'l I - 46 CI'I/.

3 lacc Vol. 8, No.2 REED ET AL. 393 anatomy was confirmed in 38 babies by obtaining normal postnatal echocardiographic examinations; all infants had normal physical examinations after birth. Our success rate for obtaining acceptable quality recordings was 96% for the mitral valve and 91% for the tricuspid valve. Peak flow velocities. The AlE ratio decreased significantly as gestational age advanced, from 1.56 ± 0.06 to 1.22 ± 0.04 across the tricuspid valve (p < 0.001) and from 1.55 ± 0.04 to 1.22 ± 0.06 across the mitral valve (p < 0.001) (Table I). Peak flow velocity during early diastole across the tricuspid valve increased significantly with gestational age, from 26.3 ± 1.0 in Group 1 to 36.5 ± 1.7 in Group 4 (p < 0.001), whereas peak flow velocity during late diastole did not change. Peak flow velocity during early diastole across the mitral valve did not change significantly, while peak flow velocity during late diastole decreased from Group 2 to Group 4 (45.8 ± 1.3 to 34.2 ± 2.7 cm/s, p < 0.001). In 91 fetuses with both tricuspid and mitral flow velocities measured at the same examination, peak flow velocity duror K.L.R.). Estimates of the angle of incidence between the direction of blood flow and the Doppler beam varied by less than 10 in 10 fetuses examined sequentially by two examiners who alternated tricuspid and mitral interrogation of the fetal heart at the same sitting in all 10 fetuses. Only flow velocity tracings obtained at angles of less than 30 were used in data analysis. Fetal heart rate was measured from the E to E peak interval and averaged for the three beats used. All velocity measurements were performed by one examiner (K.L.R.). Repeated measurements of peak flow velocities from the same Doppler page print outputs (intraobserver variability) varied less than 3 cm/s for both early and late diastolic peak flow velocities. For purposes of analysis of changes in filling patterns as a function of gestational age, we attempted to measure the peak flow velocities during atrial contraction and peak flow velocities during early diastole in four groups of fetuses: Group I, 17 to 24 weeks of gestation (n = II); Group 2, 25 to 30 weeks (n = 43); Group 3, 31 to 36 weeks (n = 47); and Group 4, 37 to 42 weeks (n = 12). The ratio of A to E (AlE ratio) was calculated in these four groups and compared (Table I). Statistics. Statistical analysis was performed using analysis of variance with significance at the 0.05 level, regression analysis and Student's t test for paired and unpaired values. Results are presented as mean ± SEM. Results One hundred twenty fetuses between 17 and 42 weeks' gestational age were examined. Some fetuses were examined more than once before birth (see later). Normal cardiac Table 2. Comparison of Tricuspid and Mitral Valve Peak Flow Velocities (cm/s) During Early and Late Diastole in the Human Fetus (n = 91) TV E 34.9 ± 0.6* TVA 46.1 ± 0.9* MVE 29.8 ± 0.6* MVA 40.9 ± 0.9* *p < Both tricuspid valve and mitral valve velocities were obtained during the same examination. Abbreviations as in Table I. Table 1. Changes in Early (E) and Late (A) Diastolic Velocities and AlE Ratios Across the Tricuspid and Mitral Valves in the Human Fetus During Gestation Group I Group 2 Group 3 Group 4 (17 to 24 weeks) (25 to 30 weeks) (31 to 36 weeks) (37 to 42 weeks) (n = II) (n = 43) (n = 47) (n = 12) P Value TV AlE 1.56 ± 0.06* 1.39 ± ± ± 0.04* *< No MVA/E 1.55 ± 0.04* 1.47 ± ± ± 0.06* *< No. II TV E (cmis) 26.3 ± 2.0* 34.4 ± ± ± 1.7* *< No TV A (cm/s) 40.9 ± ± ± ± 2.5 NS MV E (cmis) 26.3 ± ± ± ± 1.5 NS No. II MV A (cmis) 40.7 ± ± 1.3* 39.9 ± ± 2.7* *< AlE = ratio of peak flow velocity during atrial contraction (A) to peak flow velocity during early diastole (E); MV = mitral valve; No. = number of examinations in which tricuspid or mitral valve flow was successfully recorded; TV = tricuspid valve.

4 394 REED ET AL. JACe Vol. S. No.2 August J986:391-5 ingearly diastole was higheracross the tricuspid valve (34.9 ± 0.6 crn/ s) than across the mitral valve (29.8 ± 0.6, p < 0.00(1) (Table 2). Similarly, when peak flow velocities during late diastole were compared, the late diastolic tricuspid velocities(46.1 ± 0.9 cmls) were higher than mitral velocities (40.9 ± 0.9 cmls) (p < ). The effects ofheart rate did not seemtoexplainthe results because heart rate did not vary with gestational age in the fetuses we studied (the correlation coefficient of heart rate versus estimated gestational age was r = 0.15). Serial studies. A subset of this group underwent study on more than one occasion. A total of 12 fetuses had more than one study in this series, accounting for 26 studies in total, Ten had studies when they were in Groups 2 and 4. One fetushad studies when it was in Groups 1and 3; another fetus was studied when it was in Group I and then in Group 4. In six of seven of the fetuses cited earlier whosetricuspid valve was studied serially at increasing gestational ages (6 to 14 weeks betweenexaminations), the AlE ratio decreased; in 9 of 12serial studies of the mitral valve, in fetuses studied again 8 to 15 weeks after the initial examination, the AlE ratio decreased with advancing gestational age. Discussion Filling patterns of right and left fetal ventricles. In this study, ultrasound techniques were used to evaluate ventricular filling velocities as a potential index of diastolic function not previously studied in human fetuses. Doppler flow velocity tracings across atrioventricular valves showed a consistent relationof A (late diastole) and E (earlydiastole) peaks. The AlE ratio was shown to be consistently above unity. In adults, this ratio for the mitral valve is usually less than 1 (13,14), but early diastolic velocities have been shown to be lowered in elderly patients and lowering of early diastolic filling velocities has been associated with decreased left ventricular compliance (13-17). Thus it appears that the human fetal ventricle fills with a pattern that suggests that during gestation ventricular compliance is lower than it is after birth (18). This is consistent with previous data (9) in fetal lambs in which the fetal ventricle appears to be less compliant when compared with the neonatal ventricle. The explanationfor these differencesin compliancehas been attributed to a difference in the relative distribution of contractile and noncontractile elements in the fetal myocardium compared with the adult ventricle (9). Changes with increasing gestational age. A decrease in the NE ratio was demonstrated in both the right and the left ventricle as gestation advanced. In the tricuspid valve. peak flow velocity in early diastole increasedwith advancing fetal age, whereas peak flow velocity during late diastole did not change. Mitral valve changes differed, with peak flows during early diastole not changing and a significant decrease in peak flow velocity during late diastole. The different contributions of the early and late filling phases of each ventricle to the decrease in AlE ratio may reflect inherent differences in ventricular function and its maturation before birth. Tricuspid versus mitral flow velocities. Tricuspid valve peak flow velocities were greater than mitral valve peak flow velocities in both early and late diastole. Higher tricuspid peak flow velocities, when combined with the greater size of the right ventricle compared with the left ventricle (I) at the level of the valve anulus, further support the concept of right heart dominance in the human fetus. Although we did not calculate mean temporal velocity for all fetuses in this study, we have demonstrated in a previous study (5) that tricuspid volume flow is greater than mitral volume flow in the normal human fetus by a ratio of 1.3 to I. Other investigators (8) have foundevidence for right heart dominance in the study of right and left ventricular outputs in the fetal lamb. Conclusions. This study used Doppler echocardiographic techniques to demonstratedifferences in ventricular filling patterns between the fetal heart and the hearts of childrenor adults. We also demonstratedsequential changes in human fetal ventricular filling patterns during gestation. We believe that Doppler techniques will probably prove to be complementary to imaging techniques for evaluating the structure and function of the fetal heart. References I. Sahn DJ. Lange LW, Allen HD, et al. Quantitative real-time crosssectional echocardiography in the developing normal human fetus and newborn. Circulation 1980;62: Allan LD, Tynan MJ, CampbellS, Wilkinson JL. Anderson R. Echocardiographic and anatomical correlates in the fetus. Br Heart J 1980;44: WladimiroffJW, McGhie J. Ultrasonic assessment of cardiovascular geometry and function in the human fetus. Br J Obstet Gynaecol 1981:88: DeVore GR, Donnerstein RL. Kleinman CS, Platt LD, Hobbins JC. Fetal echocardiography. I. Normal anatomy as determined by realtime-directed M-mode ullrasound. Am J Obstct Gynecol 1982; 144: Reed KL, Meijboom EJ, Sahn DJ, Scagnelli SA. Valdes-Cruz LM, Shenker L. Cardiac Doppler flow velocities in human fetuses. Circulation 1986;73: Versprille A. Jansen JRe. Harinck E, van Nie C1. de Ncef KJ. Functional interaction of both ventricles at birth and the changes during the neonatal period in relation to the changes of geometry. In: Longo LD, Reneau DD. eds. Fetal and NewbornCardiovascularPhysiology. Vol 1. New York: Garland STPM Press. 1978: Azancot A, CaudellTP. Allen HD, et aj. Analysis of ventricularshape by echocardiography in normal fetuses, newborns and infants. Circularion 1983:68: Rudolph AM. Heymann MA. Fetal and neonatal circulation and respiration. Annu Rev Physiol 1974;38: RomeroT. CovellJ, FriedmanWF. A comparisonof pressure-volume relations of the fetal, newborn and adult heart. Am J Physiol 1972;222:

5 lacc Vol. 8, No.2 REED ET AL Rudolph AM. Fetal circulation and cardiovascular adjustments after birth. In: Rudolph AM, ed. Pediatrics. 17th ed. New York: Appleton Century-Crofts, 1982: II. Sabbagha RE, Hughey M. Standardization of sonar cephalometry and gestational age. Obstet Gynecol 1978;52: Hohler CW, Quetel TA. Fetal femur length: equations for computer calculation of gestational age from ultrasound measurements. Am 1 Obstet Gynecol 1982;143: Dabestani A, Johnston WD, Tobis 1M, et al. Relation between Doppler transmitral diastolic flow and left ventricular filling (abstr). 1 Am Coil Cardiol 1984;3: Gardin 1M, Dabestani A, Rohan MK, et al. Noninvasive studies of ventricular filling with Doppler echocardiography: effects of aging on early and late diastolic transmitral flow (abstr). 1 Am Coli Cardiol 1984;3: Fisher DC, Voyles WF, Sikes W, Greene ER. Left ventricular filling patterns during ischemia: an echo/doppler study in open chest dogs (abstr). 1 Am Coli Cardiol 1984;5: Kitabatake A, Inoue M, Asao M, et al. Transmitral blood flow reflecting diastolic behavior of the left ventricle in health and diseasea study by pulsed Doppler technique. 1pn Circ ;46: Miyatake K, Okamoto M, Kinoshita N, et al. Augmentation of atrial contribution to left ventricular inflow in aging assessed by intracardiac Doppler flowmetry. Am 1 Cardiol 1984;53: Grenadier E, Lima CO, Allen HD, et al. Normal intracardiac and great vessel Doppler flow velocities in infants and children. 1 Am Coli Cardiol 1984;4: