Determination of intervillous flow in early pregnancy

Placenta (1997), 18, 287-293

Determination of Intervillous Flow in Early Pregnancy N. A. B. S i m p s o n a'c, C. N i m r o d a'd, R. D e V e r m e t t e a and J. F o u r n i e r b a Division of Perinatology, Ottawa General Hospital, Ontario, Canada Animal Resources Division, Health Canada, Ottawa, Ontario, Canada Paper accepted 23 September 1996

The process of placentation in the macaque has been extensively studied and fotmd to resemble closely that observed in the human. In this model, histopathologically, intervillous flow is anticipated from week 3 post-conception. We set out to document the nature and onset of intervillous flow in the macaque in vivo using colour Doppler imaging (CDI), colour Doppler energy (CDE) and pulsed-wave Doppler (PWD). Pregnant females were assessed between 15-50 days gestation (term=165 days) with an Acuson 128/XP10 high-resolution ultrasound scanner, using a 7-MHz linear array probe. The placenta, subjacent decidua and myometrium were assessed using CDI and CDE. Specific regions of flow were interrogated using PWD; the resulting flow velocity waveforms were stored and quantified using conventional Doppler indices. B-mode sonography was able to demonstrate the well-defined placental-decidual interface observed in this species; CDI and CDE clearly visualized the uteroplacental vasculature. Spiral arteries were followed to their point of discharge into the intervillous space, and PWD at these sites obtained a characteristic flow velocity waveform. The indices obtained confirmed a flow of low resistance and pulsatility throughout the gestation studied. Flow within the intervillous space was noted from day 20 of gestation. (~) 1997 W. B. Saunders Company Ltd Placenta (1997), 18, 287-293

INTRODUCTION

Development of the villous haemochorial placenta has been documented in both man and the non-human primate (\Vislocki and Streetcr, 1938; Hamilton and Boyd, 1960; Harris and Ramse.v, 1966; Ramsev and Harris, 1966; Luckett, 1974). Prelacunar, lacunar, and villous stages are used to describe the initial process of placentation following implantation (O'Rahilly and Muller, 1987). The decidual vasculature becomes progressively disrupted by the invading trophoblast over the first 3 weeks, resulting in the appearance of red blood cells within the early intervillous space. With the advent of visible heart motions during week 4, a fetal circulation is established which may be observed within both the umbilical and intraplacental vessels. There remains uncertainty as to when an effective intervillous circulation is established (Jauniaux, Jurkovic and Campbell, 1995; Moll, 1995); light microscopy in man and non-human primate has indicated that a venous component is evident from days 11 to 14 post-conception (Hamilton and Boyd, 1960; Harris and Ramsey, 1966; Enders and King, 1991), and arterial connections have been noted from days 26 to 29 (Ortmann, 1938; Hamilton and Boyd, 1960; Ramsey, 1981; Ramsey and Donner, 1988). In vivo studies, designed to observe dynamic change, have also been reported. The only Current address: Division of Obstetrics and Gynaecology, St James's University Hospital, Leeds, UK. a To whom correspondence should be addressed. 0143-40I)4/97/040287 + (17 SI2.0(1/0

early radioangiographic study reported in the human observed diffusion of dye from 6 weeks menstrual age at thc terminal aspects of the spiral arteries (Burchell, 1967), in a manner similar to that noted in later pregnancy; these appearances were confined to a limited number of spiral arteries within the first trimester. No comparable studies were published in the non-human primate. Within the last decade, a number of studies have used ultrasound to study early human placentation (Hustin and Schaaps, 1987; Jauniaux, Jurkovic and Campbell, 199l; Jaffe and Warsof, 1991; Jaffe and Woods, 1993). The findings suggest that intervillous flow is not a feature of normal pregnancy until late in the first trinaester (Schaaps and Hustin, 1988; Jaffe and Warsof, 1992; Jaffe and Genbacev, 1993), prompting a revision of our understanding of early intrauterine development. However, it has been suggested that technical limitations may have precluded accurate assessment of low flow velocities (Moll, 1995), and since these studies were performed, more advanced equipment has been developed, enabling greater sensitivity. In particular, emerging technologies have improved the sensitivity of Doppler ultrasound, the technique used in these earlier studies. Where the changing strength of the received ultrasound signal (due to differing tissue transmission of the emitted ultrasound signal) dictates the relative grey-scale image, changes in the received frequency can be analysed to provide information regarding movement (i.e. of red blood cells) within the area of interest (Sumner, 1989). (~' 1997 W. B. Saundcrs Coml~any Ltd

288

Placenta (1997), \%1. 18

F i g u r e 1. Transabdominal view of the gravid uterus: note the tx~t) placental discs, sited {)n tile d.rsal and ventral walls, and the clear demarcation of the decidual-placental bc~rdcr..-\s in all t~ther figures, the transducer is sited at the top of the screen; the ultrasound beam is directed dmvnwards (monkey #85083, 38, days gestalicm).

In this way, bh)od flow can be readily detected: either a small region can be assessed to provide quantitative data on blood flow at that site (pulsed-wave 1)oppler, PWD); or blood flow can bc represented by colour superimposed on the grey-scale image to indicate flow to or from the transducer, known as colour i)oppler imaging (CDI) (Mitchell, 1990). In this instance, the saturation or hue of the culour is a qualitative measure of velocity. The main limitation of these two techniques is angle dependence: detection of t]ow is dependent on the direction of flow rclative to the transducer; and if perpendicular, may not be appreciated at all. A morc recent technique, colour Dopplcr energy (CDE), in measuring the amplitude or strength of the received signal, assesses the relative concentration of the reflectors (i.e. the red blood cells) as opposed to their velocities, and is therefore less angle dependent, allowing for a more sensitive detection of blood flow (Rubin et al., 1994; Kremkau, 1995). It was therefore our intention in this study to use these improvements in ultrasound technology to examine the carh uteroplacental circulation; our animal model was the cynomolgus monkey (Macaca fiascicularis), a non-ht, man primate known to exhibit trophoblastic invasion and villot, s haemochorial placentation in a manner analogous to the human (Ramse.v and Harris, 1966; l,uckett, 1974).

MATERIALS AND METHODS All investigations were carried oft after ethical approval at the non-human primate breeding colorer, Health Canada. Twentyeight females wcre studied. They were caged with fertile males fi~r 5 days at mid-cycle; gestational age was calculated from the

middle day of exposure. They were st, bsequently assessed by ultrasound 15-50 days later. Each was restrained for the ultrasound examination; none required sedation. All studies were performed with an Acuson 128/XP10 high-resohttion ultrasound scanner (Acuson, Mountain View, Califi)rnia) using a 7-MHz linear array probe with CDI, CDE and PWI) capabilities. The spatial peak temporal average (SPTA) intensities assigned to the transmitted signal ranged up to 800 m \ ~7/ c m - , with a high-pass filter set at 125 Hz. In the cynomolgus monkey, the gravid uterus is located directh" beneath the anterior abdominal wall, affording excellent transabdominal views in early pregnancy (Tarantal and Hendrickx, 1988). As in most cercopithecoids, bidiseoid placentatitm is observed, such that there exists a primary (receiving the umbilical vessels) and a secondary disc (sited on the opposite uterine wall) (Myers, 1972). The arrangement is analogous to the succentvriate lobe observed in the human. In earh pregnancy, the discs are connected bv a plexus of communicating vessels running through the chorionic (coeh}mic) cavity (Wisloeki and Hartman, 1929). In each gravid female, the uterus was assessed in transverse and sagittal planes using grey scale to establish the site of each disc, where umbilical vessels were evident. CDI and CI)E were then used to assess each disc, the subjacent decidua and mvometrium. Identified regions of flow within the intervilh}us space were interrogated using pulsed wave Doppler, and the waveforms stored. In each case the wavefc}rms obtained were expressed quantitatively bv the pulsatility index [(systolic peak v c l o c i t v - e n d diastolic velocity)/time averaged veh}city[ and resistance index [(systolic peak v e l o c i t y - e n d diastolic vch~city)/systolic peak velocity].

Simpson cta].: lnlcrvillous Blood Flow

289

Figure 2. CI)I depicts tlow from the spiral artery into the intracotylcdonary space. Blue pixcls represent flow away from the transducer, red pixcls flo~ to~xartls; n()tc the c()nvolulcd nature of the spiral artery ;it the dccidual-placcntal intcrlhcc, and the discharge of its contents onward into the substance of the placenta (monkey #85083, 3N days gestation; region imaged measures 15 × 20 ram).

F i g u r e 3. CI)I': demonstrates the n~orpllologyof the xasculaturc at the dccitlual-placcntal border. The intensity of the colour corresponds to the strength t)l"the i't2cci'~ct]signal; nolo the passage of blood into the intracotylcdonary space (monkc.~ #89062, 31 days gestation; region imaged measures 211× 20 ram).

RESULTS T h e presence of a gcstational sac was the earliest indication of pregnancy (day 15), follo~ved by the appearance of the placental discs and secondary yolk sac (day lg), fetal pole (day 21), and fetal cardiac activity (day 25). T h e demarcation between placental disc and dccidua was always clearly demonstrated (Figure 1), enabling acct, rate

regional assessment of flow. C D I and C D E revealed the course of the uteroplaccntal vessels supplying the placenta, and in particular the entry of the spiral arteries into the intcrvillous space (Figures 2 and 3). T h e s e entry points appeared to be confined to the more central portions of the placental discs. C D I and C D E enabled accurate positioning of the pulsed wave l ) o p p l c r gate, which then revealed the characteristic fizatures of the flow velocity wavcform at this site (Figure 4): a diminished

290

Placenta (1997), Vol. 18

Figure 4. Flow velocity waveform obtained from the intracotyledonary space previously identified by CDI in Figure 2; note the unidirectional nature and low pulsatility of blood flow at this site (monkey #85083, 38 days gestation; vertical axis represents flow velocity in m/s, horizontal axis depicts time in 0.2 sec).

Figure 5. Flow velo~:ity waveform obtained from the uterine artery at its entry into the cervix; compare the morphology of this waveform with that of Figure 4 (monkey #85083, 20 days gestation; axes as Figure 4).

Simpson et al.: Intervillous 131cmd Fhm' and delayed systolic peak with continuous diastolic flow, contrasting with that observed in the uterine artery (Figure 5). Qtmntitative indices confirmed that flow in this region was of h)w pulsatility and low resistance. The mean pulsatility index over the gcstation studied was 0.28 (standard deviation, s.d.=0.12) and thc mean resistance index 0.25 (s.d.=0.09); there was no significant trend with gestati(mal age. The earliest determination of intervi]lous tlmv was at 20 days gestation.

DISCUSSION Using ultrasot, nd, we have been able to determine the nature and onset of flow within the intervillous space in early pregnancy. The infi)rmation provided hy grey-scale ultrasound enahled accurate localization of the placcntal-dccidual interface; the supcrhllposed .-epresentation of blood flmv hv CI)I and CDE assisted p.'ccisc placement of the pulsed-wave gate, which subsequentl.v confirmed blood flow at these sites. This systematic approach has made t, sc of the strengths of each technique to elaborate ml predictions made fi'om the light microscopic, radioangiographic and injection-corrosion studies (Ramse 3, Corner and l)oxmer, 1963; F.'cese, 1966; Freese, Ranniger and Kaplan, 1966; \Vigglesworth, 1967; Ramsey, Chez and l)oppman, 1979). These studies indicated that after a period of coiling, the spiral arteries open into a rclativeh vilhms-frcc intracotvlcdonarv space, into which the hlood tltms before dispersing laterally and superiorly vis a tergo through the interstices provided by the villi to reach the periphery of the placentone and thence to the vcnotls system. \Ve helieve that our methodology has acct, ratelv enahled visualization of blood llo~v within decidual portions of the spiral arteries and its st,bscqt,cnt course into the intracotvledonarv portion of the intervillous space. \Vc have not ohscrved thin hetween individual villi, as the resolution of our system would not permit representation of tlmv through channels of such small dimensions, l tmvcver, from the awfilahlc histological data, fhm into this space implies st,hscquent passage. In the same wa.v, radioangiography was ablc to demonstrate central points of entry from the termination of the spiral arteries, anti Ihm hetxveen individual villi, although not secn, was implicit from the 'smoke-rings" or 'dotmghnuts' subsequently observed (1;',amscy, Corner and l)onner, 1963; l"recse, Rannigcr and Kaplan, 1966). As regards the nature of intervillous floxv, we have demonstrated three defining characteristics at this site as indicated by the llmv velocity waveform: it is unidircctimaal, pulsatile, and of hm pulsatilit.v. First, it is evident that blood flowing ti-om the t, teroplacental vasculature into the intcrvillous space at this site is unidirectional, indicating that there is no imnlediate arteriovenous connection at the basal plate; rather, that there is flow onward into the st, hstance of the placenta. This is cvident from the CDI images as well as the [l(m velocity waveform obtained. The presence of decidual arteriovenous shtmting might have provided some explanation tbr the low pulsatility of flmv previously observed in spiral arteries (Jaffc and \Varso[; 1991;

291

Jauniaux, Jurkovic and Campbell, 1991; Jurkovic et al., 1991; Jaffe and Woods, 1993); however, these connections have ncver been demonstrated histologically, and the evidence from our study does not support this theory. Second, thc flow, although continuous, is cleal'lv pulsatile. There is significant diastolic flow supplementcd by periodic systolic contributions. This would be in accord with previous observations using radioangiography which arrived at a similar conch, sion later in pregnancy (Ramsey, Corner and Donner, 1963). It is evident that the element of pulsatility is not lost in early pregnancy, which may have physiological implications for effective cotvledonarv developnaent (Reynolds, 1966, 1972). The restriction of flow to the more central portions of the placenta enables each disc to remain attached to the decidt, a. Third, the tlow at this site is of Imv pulsatility. The blood is clearly subject to an ahrupt lhll in resistance, as from a clearly dcfined vascular compartment into a relatively conlpliant space. "l'lac consequences of this are again evident in the thin velocity wavcform, in that the systolic peak is delayed anti diminished in comparison with the uterine signal. This is analogous to the pulsus parvus et tardus ohscrved distal to arterial stenoses (I,afnrtune et al., 1992; Patriquin et al., 1992; Stavros et al., 1992), and indicatcs a significant thll in the resistance to tlow (Bude et al., 1994). In addition, quantitative analysis indicates blood tlmv of low pulsatility at this site, principally due to the relatively high diastolic flow, again indicating a Imv-resistance system. It is notable that the pulsatility and resistance indices remained lmv throughot, t the gestatiml studied, presumably cnahling placental developnacnt to proceed without exposure to excessively pulsatile arterial blood flmv. The onset of intcrvilhms tlmv has clearly now been shmvn to start in earh pregnancy in the primate, and complements the histological evidence previously mentioned. These studies had demonstrated a defined temporal sequence of vascular connections to the intcrvillous space: capillary, venous, arterial, so ensuring that efl'cctive drainage of the intervillous space could take place befiwc an arterial supply was established. In the monkey, the trophohklstic lacunae communicate with neighhouring capilla.'ies from day 11 post-conception, and commtmications with ncighhouring venulcs arc evident fi'om days 11 to 15 (\Vislocki anti Strcctcr, 1938; Enders and King, 1991). \Vislocki and Streetcr (1938), noting the changes in red cell morphology within the developing intervilhms space, oraleluded that from dax 22 an effective intcrvillot, s circulation was likely to have been estahlishcd, and subsequent investigators have indeed demonstrated direct arterial connections fi'om day 29 (Ramsey and Donner, 1988). In the human, the chronology of events is strikingly similar. Capillary connections have been detailed fi'om as early as 7 days (I larris and Ramsey, 1966). ('ommunication with ncighbouring venous sinusoids has been noted to be present from days 11 to 16 (Hertig, Rock and Adams, 1956; Hamilton and Boyd, 1960; Harris and Ramsey, 1966). Arterial connections are evident fi-om days 26 to 29 (Ortmann, 1938; Hamilton and Boyd, 1960; Harris and Ramsey, 1966) albeit by indirect or labyrinthine

292

pathways, although the elegant micrographs furnished by Ortmann (1938) in his stud.v of a 26-day specimen clearly show direct channels between the spiral artery and the intervillous space. It is most probahly that arterial flow is no doubt to some extent impeded by intralunainal truphoblast in the carl.v stages, but fi'om our t)bservations, it seems evident that in thc more centrally located spiral arteries, effective flow exists from wcck 4 of gcstatit)n. The trophoblastic shell is intact in tbc peripheral portion of the placental disc anti this allows fur the anchoring process to hc maintained. The development of the intcrvillous flow is more advanced ccntt'allv as the maternal vessels perforate the tropht)blastic shell and merge with the developing lacunae within it to form the primitive circttlation. It is in this location that the CI)I is first seen as the peripheral portions fl'ee of an cffcctivc intcrvillous circulation. In the human, previous

Placenta (1997), Vol. 18

ultrasound studies have not noted intervillous flow in normal pregnancies until the end of the first trinacster. It is notable that in this study, certain technical improvements permitted optimal visualization of the developing placenta: a linear array probe; the use ofa 7-MHz transmitting frequency and high SPTA intensities; the close proximity of the site of interest to the probe; the use of CDI and CDE, and PWD. Such advantages are not readily available in the study of ongoing human pregnancy, where regulations limit the use of Doppler ultrasound in early gestation. However, wc fccl that there are such close associations in placcntation between tbc two spccics that we should expect intervillous flow to he present at an early stage of normal pregnancy; it is likely that under current technical limitations the use of echo-contrast agents (Goldbcrg, I,iu and Forsbcrg, 1994) may well assist in its determination in earls" human gestation.

ACKNOWLEDGEMENTS .'\cknowlcdgcmcr~ts arc duc t~) the :\cusort (:orporation fi)r tficir loan ill" the 12S/XF'10 higfi-rcsolution ultrasound scanner used in this stud)'.

REFERENCES Bride, R. ()., Rubin, J. M., F)latt, J. F., F e c h n e r , K. P. & .&dlcr, R. S. (1994) Pulsus tardus: its cause and potential limitations in detection i)l arterial stcnosis. Ra,li, logv, 1911, 779-784. Burcl'tell, R. C. (1967) Arterial bhlod Ihl'~ into tltc l')unaan irlter~ilh)us space. .Imcrican .7o,rmtl ~!/"Obstetrics a,d (,'.),,ccr~/og),, 98, 3113-311. Endcrs, A. C. & King, I:l. I'. (1991) I':arl3 stages (Jl"trot')fiol')lastic in~asi(m of Ific maternal :aSCtllal" system tltlrillg inll')lantalion ill tfiC macaque arid fial')ll(m..-lm,'rtc,,1.7,ttr,,tl ~¢[.'lnat,mO,, 192, 329-346. Frccse, U. E. (1<)66)The fetal-maternal circulation ()f the i')laccnta: fiistom(~rpfiologic, l')lastllitl inicctiort, and x-ray cincmatogral')fiic studies on human i')laccntas..-Imert~'an .7o,rnal *~/'Ohst¢trtcs am/(~l,necologl,, 94, 354-3611. l:reese, U. E., R a n n i g e r , K. & Kal')lan, H. ( 1 9 6 6 ) T h e fctal-matcrnal circulation t)l" tfic r)laccnta: an x-ray cinctr~atogral~l~ic study ()f pregnant rficsus monkeys..-Imcr
Jurkovic, D., J a u n i a u x , E., Kurjak, A., H u s t i n , J., C a m p b e l l , S. & Nicolaides, K. H. (1991) Transvaginal color l)opplcr assessment of tfic utcroplacental circulation in carl)" pregnancy. Ohstetrics and (~l',ec,hv~O', 77, 365-369. K r e m k a u , F. W. (1993) Contcml)orar.v uhrasound tccfinolog.~, l ¥traso,ml ill Obstetrics tlllJ (~J,tl¢(ll[o,~J¢, 6, 233-236. L a f o r t u n e , ~XI., P a t r i q u i n , It. B., D e m c u l c , E., Trinh, B-C., D u f r e s n c , M. P., L e g a u l t , L. & R a y m o n d , J. 11<)92) Renal arterial stenosis: slowed systole in the downstream circulation--experimental study in thDgs. R,t,li,hJ,~v, 184, 475-478. Luckett, W. P. 11974) (;omparativc dcvelopnlcnt and evolution of the placenta in primates. Co,tributio,s to Primatoh,A},, 3, 142-234. Mitchell, D. G. 11990) (hdor Doppler imaging: principles, limitati~ms, and artifhcts. Radt,h~
Simpson ctal.: Intervillous Blood Flow Rubin, J. M., Bude, R. O., Carson, P. L., Bree, R. L. & Adler, R. S. (1994) l)nwer l)oppler US: a potentially useful alternative to mean frequency-based color Doppler US. Radioh~g),, 190, 853-856. Schaaps, J=P. & Hustin, J. (1988) In vivo aspect of the maternaltrophoblastic border during the first trimester of gestation. Trophoblasl Research, 3, 39-48. Stavros, A. T., Parker, S. H., Yakes, W. F., Chantelois, A. E., Burke, B. J., Meyers, P. R. & Schenck, J. J. (1992) Scb,'mcntal stcnosis of the renal artery: pattern recognition of tardus and parvus abnormalities with duplex sonography. Radiol,g),, 184, 487-492. S u m n e r , D. S. (1989) Objective diagnostic techniques role of the vascular laboratory. In l'ascuhlr Sur¢eo, (I'd.) Rutherfi)rd, R. B. pp. 41-60. Philadelphia: W. B. Saunders.

293 T a r a n t a l , A. F. & Hendrickx, A. G. (1988) Use of ultrasound for karl)" pregnancy detection in the rhesus and cynomolgt,s macaqt, c (Macaca mulatta and Macaca fhscicularis). Journal ~!["Medical Primatr)hJgr, 17, 1(15-112. Wigglesworth, J. S. (1967) Vascular organization of the human placenta. Nature, 216, 112(I-1121. Wislocki, G. B. & H a r t m a n , C. G. (1929) On tile placcntation of a macaque (.\lacacus rhesus) with observations on the origin of the bh)od constituting the placental sign. Bulh'tin ~fth('Johns Hopkins llospihd, 44, 165-185. Wislocki, G. B. & Streeter, G. L. (1938) On the placcntation of the macaque (Macaca mtdatta), fi'om the time of implantatlnn until tile formation of the definitive placenta. Contributions t, Embo,ohJ¢),, 27, 1-66.