The presence of an acardiac fetus in a twin pregnancy is a rare event, occurring in one of every 35,000 pregnancies.1 However, the morbidity and mortality are not negligible; cardiac failure in the normal twin and the risk of premature delivery caused by polyhydramnios and/or rapid growth of the acardiac mass are common. It is now known that circulatory failure of the normal twin derives from the existence of arterioarterial and venovenous anastomoses within the placenta that allow retrograde perfusion of the acardiac twin by blood coming from the normal twin.2 This peculiar hemodynamic condition has been called twin reversed arterial perfusion sequence. In view of the potentially lethal cardiovascular complications for the normal twin, invasive maneuvers have been advocated, either to interrupt perfusion of the mass3–8 or to remove the acardiac mass from the uterine cavity.9–11 These procedures carry substantial risks, the most frequent being induction of premature labor. Yet, not all twin pregnancies with a twin reversed arterial perfusion sequence have a dismal outcome; the mortality rate is reported to be between 50% and 70% for the normal twin.12 Ideally, invasive interventions should be preventive measures reserved for selected cases identified as being more at risk of cardiocirculatory compromise and secondary hydrops fetalis. However, criteria for the identification of this group of fetuses do not exist.
We therefore reviewed our cases in search of Doppler echocardiographic variables that could help identify fetuses that would benefit from invasive maneuvers before evidence of hydrops. Our initial hypotheses were the following: the lower the vascular resistance offered by the acardiac mass, the greater the risk of cardiac failure and hypoxemia in the normal twin resulting from a steal phenomenon of blood destined primarily to the placenta; the degree of circulatory overload imposed on the normal fetus should be translated by a hyperdynamic heart with a proportional increase in the systolic shortening of myocardial fibers and significant cardiomegaly; and finally, in the acardiac mass, the growth profile, the number and size of cysts, and the presence or absence of a primitive heart with some autonomous function could influence the outcome.
Materials and Methods
Between January 1990 and July 1997, twin reversed arterial perfusion sequence was diagnosed in 10 pregnancies evaluated in our Fetal Cardiology Unit. The diagnosis was obtained by Doppler signal documenting retrograde arterial flow within the cord of the acardiac mass. One patient was lost to follow-up after only one visit. Most cases were evaluated serially with a mean of 4.4 visits (range 1–8).
Videotape recordings of echocardiograms of the nine fetuses for whom outcome data were available were reviewed. The ultrasound studies were conducted with a Hewlett-Packard Sonos–100 (Andover, MA) before 1991 and an Acuson 128 XP–10c (Mountain View, CA) echocardiograph thereafter. The following data were collected for the normal fetuses: cardiothoracic ratio and left ventricular shortening fraction. The acardiac mass was evaluated for maximal length, presence and size of cysts, and the presence of a pulsatile primitive heart. The pulsatility index (PI) of the umbilical arteries (UA) of both twins was also measured. The PI was obtained from the ratio of maximal systolic flow velocity—diastolic velocity divided by mean velocity.13 The ratio between the PI of the acardius and that of the normal fetus was calculated. The shortening fraction was calculated from measurements made on M-mode tracings of the short-axis view of the ventricles using the formula: left ventricular diastolic dimension—left ventricular systolic dimension divided by left ventricular diastolic dimension. The normal range of the shortening fraction during fetal life has been shown to be between 25% and 35%.14
The cardiothoracic ratio was calculated as the greater cardiac circumference over the circumference of the thorax measured on its transverse view. Cardiomegaly was defined as a ratio greater than 0.55.15 The evaluation of cysts was subjective and rated on a scale of 6, with 1 representing very small or no cysts and 6 indicating extensive cysts (unique, large cysts or multiple smaller cysts). Finally, the maximal length of the acardiac mass was taken as an approximate measure of its growth. The outcome of each pregnancy was reviewed. Adverse outcome for the normal twin was defined as death, cardiac failure, or delivery before 30 weeks' gestation for reasons related directly to the presence of the acardiac mass. Cardiac failure was diagnosed when cardiomegaly was associated with an abnormal decrease in the shortening fraction (below 25%)14 or with the presence of pericardial, pleural, or peritoneal effusion.
The patients were divided into two groups according to their final outcome (good or adverse). To assess differences between both groups, continuous variables were analyzed by the Mann-Whitney U test, whereas the χ2 test was applied to discrete data. Emphasis was placed on observations made at the first visit. Statistical significance was set at P < .05.
Table 1 presents demographic data on the 9 patients for whom follow-up data were available. Two fetuses died in utero, both at 22 weeks' gestation (cases 1 and 2). There were two neonatal deaths (cases 3 and 4), both of whom had been delivered by cesarean for advanced cardiac failure, at 26 and 31 weeks' gestation, respectively. Case 3 died at 1 week of age from multiple organ failure, including the heart. Case 4 died at 4 days of life after a course that was complicated by cardiocirculatory failure. The five other cases had a favorable outcome. Case five was delivered by cesarean at 26 weeks' gestation because of prolonged rupture of the membrane and suspicion of chorioamnionitis; the neonate was discharged from the hospital 11 weeks after birth in good health. Cases 6 and 7 were delivered spontaneously at 36 and 37 weeks' gestation, respectively. In case 7, spontaneous disappearance of flow velocity signals in the cord of the mass was documented at 30 weeks' gestation with subsequent involution of the mass. Patient 6 had a single UA, right renal agenesis, and some vertebral anomalies; the latter required multiple surgeries later in life. Patient 7 was healthy. Both neonates were discharged at 4 days of age. Cases 8 and 9 were delivered by cesarean at 34 and 31 weeks' gestation, respectively. The indications for delivery were a decrease in amniotic fluid together with an increasing size of the acardiac mass for case 8 and sustained bradycardia (90 beats per minute) for case 9. The postnatal outcome was favorable for both.
The variables found not to influence outcome are presented in Table 2. The cardiothoracic ratio at the first visit was generally higher in the poor outcome group but this trend did not reach statistical significance. The number and size of cysts and the presence or absence of a rudimentary heart in the mass did not contribute to assessment of the prognosis.
Table 3 enumerates the PI calculated on UAs of both normal and acardiac fetuses. In all but two cases (cases 5 and 8), the PI values of the UA perfusing the mass were lower than those of the normal twin on the first echocardiographic study. When pregnancies with favorable outcome were compared with those that ended with the death of the fetus, the average value of the PI of the acardiac mass at the first visit was significantly higher in the group with good outcome (1.31 compared with 0.91). Similarly, the PI ratio (acardius to normal twin) was significantly higher for the good outcome (mean 1.04) than for the group with adverse outcome (mean 0.71). All four fetuses with unfavorable outcome had a PI ratio (acardius to normal twin) lower than 0.80, whereas four of five fetuses with a good outcome had a ratio equal to or greater than 0.88.
Data on the left ventricular shortening fraction of the normal twin are also presented in Table 3. M-mode tracing at the first visit was available for two of the four cases with poor outcome. Left ventricular shortening fraction was either abnormally high (case 3) or at the upper limits of normal (case 4). Those cases showed a marked decrease in shortening fraction at the last echocardiographic study. All five fetuses in the group with favorable outcome had a shortening fraction within normal values at their first echocardiographic study. It increased during follow-up in all but one case (case 9). Although this fetus did not show clinical evidence of cardiac decompensation, he had to be delivered by cesarean for sustained bradycardia 1 week after the echocardiographic study.
The length of the acardiac mass at the first visit is shown in Table 3. The values for the group with adverse outcome were generally greater than for the fetuses who survived, but this difference did not reach statistical significance. The size of the mass increased at a variable rate, as expressed by postnatal weight (Table 1). Although weights were not available for the two pregnancies that ended with intrauterine death, an inverse relationship (r = −0.64) could be established between postnatal size of the mass and the PIs of the umbilical arteries at the last visit for the remaining cases. Furthermore, as shown in Figure 1, fetuses with a favorable outcome tended to have a PI ratio (acardiac to normal twin) greater than 1. This figure suggests that the lower the vascular impedance in the mass, the greater its size will be.
Many publications have described the pathologic findings in twin pregnancies with an acardiac mass.16–18 The objective of the present study was to find elements that could predict the final outcome and help in choosing between an expectant or an interventional approach in twin pregnancies with a twin reversed arterial perfusion sequence. Emphasis was therefore placed on data collected at the first visit during the second trimester of gestation.
The factors responsible for growth of an acardiac twin are unknown. It seems logical that progressive growth of the acardiac mass requires significant input from the normal twin and therefore higher cardiac workload. This point has been stressed previously.12 An intriguing finding in the present study was that in most cases, the PI of the reversed UA flow perfusing the mass was lower than that of the UA in the umbilical cord of the normal fetus, which perfuses the placenta. It is known that the arterial PI is influenced by downstream vascular impedance.19 The lower PI in the UA perfusing the mass would therefore suggest that its vascular network of arteries, capillaries, and veins offers less resistance than that of a normal placenta. One possible explanation for this surprising hemodynamic observation could be the presence of wide-open connections establishing arteriovenous fistulas within the mass, thus bypassing the high-resistance capillary network. To the best of our knowledge, gross fistulas have not been demonstrated in acardiac fetuses; however, they would explain the poor peripheral vascularization with abnormal development of the head as well as of the peripheral extremities normally described in these cases. Irrespective of the factors responsible for the peripheral vascular resistance, the present study indicates that the higher the resistance offered by the mass, the lower the risk of rapid growth, which should mean lower risk of heart failure and of morbidity for the normal twin.
Our hypothesis that the circulatory load imposed on the normal fetus would translate into an increased left ventricular shortening fraction was confirmed. All fetuses with a good outcome had a shortening fraction-within normal limits at the time of diagnosis. This observation suggests that the prognosis could be related directly to the precocity of the circulatory overload, ie, the earlier the hemodynamic consequences are observed, the higher the risk of morbidity. In case 3, there was a sudden decrease in shortening fraction while the PI of the acardiac mass remained low. The poor outcome of this case suggests that such an observation announces impending death by terminal cardiac failure. One can also interpret a decrease in shortening fraction from abnormally high values to normal levels while the PI for the acardius remains low (case 9) as a sign of impending cardiac failure. This particular aspect of myocardial function has been noted previously in pre-term infants with widely patent ductus arteriosus20 and is common to all conditions associated with low systemic impedance.
In two previous case reports,21,22 we hypothesized that evidence of a rudimentary heart with some autonomous circulation within the mass could decrease the workload of the normal heart. The present data do not support that assumption. There were six acardiac masses with evidence of a pulsatile rudimentary heart. Pregnancy ended with death of the normal twin in three of them, two with heart failure. Therefore, the presence or absence of a rudimentary heart in the acardiac twin does not seem to be an independent prognostic factor.
The rarity of this anomaly explains the small number of cases reported here and does not permit definitive conclusions to be reached concerning criteria that can be used for prognostic assessment. Based on the present study, however, the following elements can be proposed as preliminary criteria for an expectant approach instead of invasive maneuvers in pregnancies with a twin reversed arterial perfusion sequence: a normal left ventricular shortening fraction at the first visit during the second trimester, a high PI (> 1.3) in the UA perfusing the acardiac mass or a PI ratio close to 1 or higher than 1, and slow growth of the mass. A prospective multicenter study to collect data on more cases would help provide more information about treating these high-risk pregnancies.
1. James WS. A note on the epidemiology of acardiac monsters. Teratology 1977;16:211–6.
2. Van Allen MI, Smith DW, Shepart TH. Twin reversed arterial perfusion (TRAP) sequence: A study of 14 twin pregnancies with acardius. Semin Perinatol 1983;7:285–93.
3. Porreco RP, Barton SM, Haverkamp AD. Occlusion of umbilical artery in acardiac, acephalic twin. Lancet 1991;337:326–7.
4. McCurdy CM Jr, Childres JM, Seeds JW. Ligation of the umbilical cord of an acardiac-acephalus twin with an endoscopic intrauterine technique. Obstet Gynecol 1993;82:708–11.
5. Quintero RA, Reich H, Puder KS, Bardicef M, Evans MI, Cotton DB, et al. Brief report: Umbilical-cord ligation of an acardiac twin by fetoscopy at 19 weeks of gestation. N Engl J Med 1994;330:469–71.
6. Sepulveda W, Bower S, Hassan J, Fisk NM. Ablation of acardiac twin by alcohol injection into the infra-abdominal umbilical artery. Obstet Gynecol 1995;86:680–1.
7. Foley MR, Clewell WH, Finberg HJ, Mills MD. Use of the Foley Cordostat grasping device for selective ligation of the umbilical cord of an acardiac twin: A case report. Am J Obstet Gynecol 1995;172:212–4.
8. Rodeck C, Deans A, Jauniaux E. Thermocoagulation for the early treatment of pregnancy with an acardiac twin. N Engl J Med 1998;339:1293–5.
9. Ginsberg NA, Applebaum M, Rabin SA, Caffarelli MA, Kuuspalu M, Daskal JL, et al. Term birth after midtrimester hysterotomy and selective delivery of an acardiac twin. Am J Obstet Gynecol 1992;167:33–7.
10. Robie GF, Payne GG Jr, Morgan MA. Selective delivery of an acardiac, acephalic twin. N Engl J Med 1989;320:512–3.
11. Fries MH, Goldberg JD, Golbus MS. Treatment of acardiac-acephalus twin gestation by hysterotomy and selective delivery. Obstet Gynecol 1992;79:601–4.
12. Moore TR, Gale S, Benirschke K. Perinatal outcome of forty-nine pregnancies complicated by acardiac twinning. Am J Obstet Gynecol 1990;163:907–12.
13. Sonesson SE, Fouron JC, Drblik SP, Tawile C, Lessard M, Skoll A, et al. Reference values for Doppler velocimetric indices from the fetal and placental ends of the umbilical artery during normal pregnancy. J Clin Ultrasound 1993;21:317–24.
14. DeVore GR, Siassi B, Platt LD. Fetal echocardiography. IV: M-mode assessment of ventricular size and contractility during the second and third trimesters of pregnancy in the normal fetus. Am J Obstet Gynecol 1984;150:981–8.
15. Paladini D, Chita SK, Allen LD. Prenatal measurement of cardio-thoracic ratio in evaluation of heart disease. Arch Dis Child 1990;65:20–3.
16. Nerlich A, Wisser J, Draeger A, Nathrath W, Remberger K. Human acardiac anomaly: A report of three cases. Eur J Obstet Gynaecol Reprod Biol 1990;38:79–85.
17. Groeninghen JC, Franssen AMHW, Willemsen WNP, Nijhuis JG, Puts JJG. An acardiac acephalic monster. Eur J Obstet Gynaecol Reprod Biol 1985;19:317–25.
18. Gewolb IH, Freedman RM, Kleinman CS, Hobbins JC. Prenatal diagnosis of a human pseudoacardiac anomaly. Obstet Gynecol 1983;61:657–62.
19. Trudinger BJ, Stevens D, Conelly A, Hales JR, Alexander G, Bradley L, et al. Umbilical artery flow velocity waveforms and placental resistance: The effect of embolization of the umbilical circulation. Am J Obstet Gynecol 1987;157:1443–8.
20. Baylen B, Meyer RA, Korfhagen J, Benzing G, Bubb ME, Kaplan S. Left ventricular performance in the critically ill premature infant with patent ductus arteriosus and pulmonary disease. Circulation 1977;55:182–8.
21. Fouron JC, Leduc L, Grignon A, Maragnès P, Lessard M, Drblik SP. Importance of meticulous ultrasonographic investigation of the acardiac twin. J Ultrasound Med 1994;13:1001–4.
22. Pavlova M, Fouron JC, Proulx F, Lessard M. Importance de l'observation intra-utèrine d'une circulation autonome rudimentaire chez un jumeau acardiaque [Importance of intrauterine diagnosis of rudimentary autonomic circulation in an acardiac twin]. Arch Mai Cceur Vaiss 1996;89:629–32.