Growth pattern of twins of different chorionicity evaluated by sonographic biometry

OBJECTIVE: To determine differences between growth patterns of monochorionic and dichorionic twins, and of concordant and discordant twins of both chorionicities. METHODS: We studied 70 cases of concordant twins (24 monochorionic, 46 dichorionic) and 45 cases of discordant twins (25 monochorionic, 20 dichorionic). In each case, growth was measured longitudinally by ultrasound biometry and the growth pattern was depicted. RESULTS: There were no differences in incremental growth between concordant monochorionic and dichorionic twins. The growth curve of concordant twins of both chorionicities was almost the same as that of a singleton until 34 weeks' gestation. However, in the discordant twins, the growth of the larger twin matched the growth curve of a singleton or concordant twin, but the growth of the smaller twin gradually decreased to the range of growth restriction. The growth curves for monochorionic discordant twins appeared to be representative of two groups, one of which had onset of discordancy before 24 weeks.CONCLUSION: It is clinically important to determine chorionicity early in twin pregnancies, to calculate the percentage of discordancy between the fetuses, and to examine longitudinal fetal growth curves in each chorionicity.     

Diamniotic twin pregnancies with a single placental mass; prediction of chorionicity at 11 to 14 weeks of gestation

OBJECTIVE: To assess the accuracy of transvaginal ultrasound (US) in the prediction of the chorionicity of diamniotic twin pregnancies with a single placental mass at 11 to 14 weeks of gestation. METHOD: From June 2006 to April 2007, we determined chorionicities by depiction of the amnion and chorion at the membrane-to-placental interface using transvaginal US. Pregnancies were classified as monochorionic when two layers of the amnion were identified as dichorionic when either one layer of the chorion and two layers of the amnion or one layer of the chorion and one layer of the amnion were seen. RESULTS: In 65 out of 70 (92.9%) diamniotic twin pregnancies with a single placental mass, we were able to determine the chorionicity by depiction of the amnion and chorion at the membrane-to-placental interface using transvaginal US. The predictive accuracy was 100% (95% confidence interval: 92-100%) for 52 twin pregnancies considered to be dichorionic by transvaginal US and 100% (95% CI: 73-100%) for 13 twin pregnancies considered to be monochorionic. CONCLUSION: The chorionicity of diamniotic twin pregnancies with a single placental mass can be reliably predicted by transvaginal US depiction of the amnion and chorion at 11 to 14 weeks of gestation.     

Prediction of chorionicity in twin pregnancies at 10-14 weeks of gestation

Objective To examine the accuracy of sonographic determination of chorionicity in twin pregnancies at 10–14 weeks of gestation.
Design Prospective study on the sonographic prediction of chorionicity at 10–14 weeks of gestation.
Participants During a 30 month period, from October 1997 to May 2000, 165 women attending the departments of fetal medicine or ultrasound.
Methods Sonographic criteria used in the diagnosis of chorionicity were the number of placental sites, the lambda (λ) and T signs and the thickness of the inter-twin membrane. The diagnosis of chorionicity was made at the time of the ultrasound examination using all these features and subsequently compared with the postnatal diagnosis, confirmed either by placental histology or discordancy in infant sex.
Results In 150 cases with confirmation of chorionicity following delivery, 116 were postnatally classified as dichorionic and 34 monochorionic. Prenatal ultrasound examination correctly identified chorionicity in 149 (99.3%) cases. The most reliable indicator for dichorionicity was a combination using the λ sign or two separate placentae with a sensitivity and specificity of 97.4% and 100%, respectively. The most useful test in predicting monochorionicity was the T sign with a sensitivity of 100% and specificity of 98.2%. Measurement of the inter-twin membrane thickness was a less reliable indicator where the sensitivity for dichorionicity and specificity for monochorionicity was only 92.6%.
Conclusions Ultrasound examination of twin pregnancies at 10–14 weeks of gestation predicts chorionicity with a high degree of accuracy using a combination of the number of placentae, λ and T signs and inter-twin membrane thickness. All hospitals should encourage departments providing ultrasound services to undertake chorionicity determination when examining women with twin pregnancies at this gestation.     

Ultrasound in twins: Dichorionic and monochorionic

The objective of this review is to assess the evidence that supports the use of ultrasound in twin pregnancies. Although many of the indications for obstetric ultrasound are the same in both singleton and multiple gestations, there are special considerations as well as unique conditions in twins that require additional imaging studies. The reasons for ultrasound in twins include pregnancy dating, determination of chorionicity, nuchal translucency assessment, anatomical survey, placental evaluation, cervical length assessment, routine fetal growth, and serial surveillance of pregnancies complicated by anomalies, cervical shortening, fetal growth disturbances, and amniotic fluid abnormalities. Twins with monochorionic placentation require heightened scrutiny for monoamnionicity, conjoined twins, twin reversed arterial perfusion (TRAP) syndrome, twin–twin transfusion syndrome, unequal placental sharing with discordant twin growth or selective intrauterine fetal growth restriction (IUGR), twin anemia–polycythemia sequence (TAPS), and single fetal demise. Ultrasound is essential for the detection and management of conditions that can complicate dichorionic and monochorionic twin pregnancies.     

Prenatal diagnosis of zygosity by fetal DNA analysis,a contribution to the management of multiple pregnancies. A series of 31 cases

OBJECTIVE: To evaluate the contribution of prenatal diagnosis of zygosity by fetal DNA analysis to the management of multiple pregnancies. METHODS: Between March 1999 and March 2000, 31 same-sex multiple pregnancies, were referred to our Institute during their second trimester to have amniocentesis done. Fetal DNA variants were studied in addition to karyotype. The main indications for amniocentesis were fetal growth discordance, fetal malformations or selective pregnancy termination. Zygosity results were compared to the first-trimester ultrasound diagnosis. RESULTS: 21 out of the 31 multiple pregnancies were dizygous, and 10 out of the 31 pregnancies were monozygous. First-trimester ultrasound provided information on chorionicity only in 24 out of the 31 cases (77%). Of these 24 reports, 21 proved to be correct (87.5%). Moreover, in 2 cases among 14 pregnancies with growth discordance, DNA analysis ruled out a twin-to-twin transfusion syndrome, while ultrasound could not yield a conclusion. In 2 twin pregnancies with fetal malformations affecting one of the fetuses, DNA analysis made it possible to assess the risk for the other twin. In the 5 cases leading to selective termination, prenatal diagnosis of zygosity had a decisive value for the surviving fetus. CONCLUSION: Fetal DNA analysis is a simple and highly useful test for prenatal diagnosis of zygosity in the management of complicated multiple pregnancies.     

Discordant lower urinary tract obstruction in early twin gestations: management and outcome

OBJECTIVE: To report our experience with the management of twin pregnancies discordant for lower urinary tract obstruction. METHODS: Cases of twin pregnancies discordant for lower urinary tract obstruction were identified from our fetal medicine database. Information on ultrasonographic findings, antenatal course, pregnancy complications, and perinatal outcome was obtained by reviewing medical records or contacting the referring obstetricians. RESULTS: Five twin pregnancies discordant for lower urinary tract obstruction were diagnosed between 11 and 15 weeks of gestation. There were 3 dichorionic and 2 monochorionic pregnancies (1 diamniotic and 1 monoamniotic). The dichorionic pregnancies were managed conservatively, resulting in a pregnancy loss of both twins in 1 case, a single fetal death at 29 weeks in 1 case, and an early neonatal death due to lung hypoplasia of the affected twin in 1 case. On the other hand, both monochorionic twin pregnancies were managed with serial vesicocenteses. In both cases, the prenatal course was complicated, 1 by premature rupture of the membranes and the other by cord entanglement, requiring delivery at 29 and 31 weeks, respectively. Among the 4 continuing pregnancies with complete perinatal outcome, none of the affected twins survived, and the structurally normal twins were delivered between 29 and 36 weeks and discharged from the hospital in good condition. CONCLUSION: Twin pregnancies discordant for lower urinary tract obstruction are at high risk of perinatal death and premature delivery. Prenatal intervention seems not to be associated with an improved perinatal outcome of the affected twin, but it may be beneficial in selected cases to attain viability of the unaffected twin.     

Management of twin pregnancies discordant for anencephaly

Objective To examine options of management and outcome of twin pregnancies discordant for anencephaly.
Design Retrospective study.
Setting Research Centre for Fetal Medicine.
Population Twenty-four twin pregnancies discordant for anencephaly.
Methods A computer search was made of our database for twin pregnancies discordant for anencephaly. The data were reviewed for gestation at presentation, chorionicity, management and pregnancy outcome.
Main outcome measures Pregnancy outcome in relation to chorionicity and management.
Results There were 13 dichorionic and 11 monochorionic twin pregnancies discordant for anencephaly. In the dichorionic group five pregnancies had selective fetocide at 17 to 21 weeks; one pregnancy resulted in spontaneous abortion but in the others a healthy infant was born at a median gestation of 37 weeks. The other eight dichorionic pregnancies were managed expectantly, but three developed polyhydramios at 26 to 30 weeks; in one case amniodrainage was performed and in another selective fetocide was carried out. In this group the median gestation at delivery was 35 weeks. All 11 monochorionic pregnancies were managed expectantly and in three there was intrauterine death of both fetuses. In the other eight cases the normal twin was liveborn at a median gestation of 34 weeks; in four of these pregnancies polyhydramnios developed and two were managed by amniodrainage.
Conclusions In monochorionic pregnancies, expectant management is associated with a high rate of intrauterine lethality of the normal twin. In dichorionic pregnancies selective fetocide in the second trimester prevents the development of polyhydramnios and is associated with a lower risk of preterm delivery but can cause miscarriage.     

Twin pregnancy: the role of ultrasound in management

Determination of chorionicity is one of the most important issues in the management of twin pregnancy. Modern ultrasound equipment has made it possible to accurately assess placentation already in the first trimester with the lambda sign. With regard to prenatal diagnosis, it is important to know the chorionicity in order to calculate the risk of chromosomally abnormal fetuses. Accurate chorionicity offers the obstetricians the opportunity to observe the monochorionic twins more intensively than is required for twins with dichorionic placentation. This review gives an update of the state of the art for clinicians caring for twin pregnancies.     

The sonographic diagnosis of chorionicity

The differentiation between mono- and dichorionic placentation in twin pregnancies is of clinical importance because of the significant difference in perinatal morbidity and mortality between the two, and the increased surveillance indicated in monochorionic gestations. Application of ultrasonography has enabled very precise prenatal determination of chorionicity. While this is best performed in the first trimester when accuracy approaches 100%, even in the third trimester, using a composite cascade of available sonographic features, accuracy has been reported to approach 97%. While two clearly separate placentae or discordant fetal gender conform to dichorionicity, in most twin pregnancies other features need to be assessed to determine chorionicity. The presence of the 'lambda' or the 'T' sign in the presence of a single placenta, best determined in the first trimester, is the most reliable indicator of chorionicity, with measurements of the inter-twin membrane thickness and counting of the membrane layers being less reliable. In this article, we review the sonographic features that help in the accurate depiction of chorionicity.     

Does chorionicity or zygosity predict adverse perinatal outcomes in twins?

OBJECTIVE: The purpose of this study was to evaluate chorionicity and zygosity as risk factors for adverse perinatal outcomes in twins. STUDY DESIGN: A population-based, retrospective cohort study was conducted of all twin deliveries in Nova Scotia, Canada, from 1988 to 1997. Chorionicity was established by histologic examination. Zygosity was determined by chorionicity, sex, and infant blood group. Three groups were established: monochorionic/monozygotic twins, dichorionic/dizygotic twins, and dichorionic/majority monozygotic twins. RESULTS: Outcomes from 1008 twin pregnancies were analyzed. Monochorionic/monozygotic twins had lower mean birth weights compared with dichorionic/dizygotic twins. Rates of perinatal mortality of at least 1 twin were significantly higher among monochorionic/monozygotic twins relative to dichorionic/dizygotic twins (relative risk, 2.5; 95% CI, 1.1-2.5). Dichorionic/majority monozygotic twins had similar perinatal outcomes compared with dichorionic/dizygotic twins. CONCLUSION: Monochorionicity increases the risk of adverse perinatal outcome, whereas the effect of zygosity is less clear. Because chorionicity can be determined by prenatal ultrasound scanning, this information should be considered in the prenatal care of twin pregnancies.     

Antepartum management of twin gestation: ultrasound

1. The use of ultrasound may improve perinatal outcome in twin pregnancy. 2. Singleton growth curves for FL, BPD, and AC may be used to assess fetal growth prior to 28 weeks gestation. 3. After 28 weeks gestation, specific twin growth curves should be used. However, increased morbidity may be seen when fetal weight falls from the normal singleton predictive curves. 4. Amnionicity and chorionicity frequently can be determined. This should be attempted early in the second trimester so that appropriate monitoring can be offered for the monoamnionic and monochorionic gestations. 5. Serial assessment of fetal growth should be performed. In the authors' institutions, growth scans are performed in twins every 2 weeks after 24 weeks gestation. 6. Evidence on ultrasound examination of discordant growth suggests either fetal growth retardation or TTS. In the presence of a dichorionic placentation, growth retardation of one twin is the usual diagnosis. If placentation is monochorionic, it is difficult to differentiate between growth retardation and TTS. The presence of polyhydramnios, a stuck twin, or fetal hydrops are suggestive of TTS. 7. The value and safety of cordocentesis and the management of twin pregnancy and diagnosis of TTS have not been established. 8. Doppler ultrasound may be useful in the assessment of the IUGR twin fetus. The use of Doppler in the identification of TTS is not yet clear.     

Perinatal outcome and neurological follow up of the cotwins in twin pregnancies complicated by single intrauterine death

Objective

To review the outcome of twin pregnancies complicated by single fetal intrauterine death (IUD) managed at our Centre and to evaluate the neurological follow up of the surviving cotwins.

Study design

Twenty-three twin pregnancies (10 dichorionic and 13 monochorionic diamniotic) complicated by IUD in the II or III trimester were seen at our Centre during the study period (2001–2006). All patients were managed conservatively unless non-reassuring signs of fetal well-being were present at ultrasound examination or CTG after 28 weeks, suggesting immediate delivery. Serial scans after the diagnosis of single death were performed and, in addition, eight monochorionic twin pregnancies underwent prenatal MRI in order to identify the presence of cerebral lesions in the survivors. Live born surviving cotwins underwent neurological follow up.

Results

In the monochorionic group one cotwin died in utero and one in the neonatal period with a perinatal survival rate of 83.4% (10/12) (excluding one case who opted for termination of pregnancy); in the dichorionic group perinatal survival rate was 100%. In all monochorionic cases there were no signs of ischemic brain lesions in the surviving cotwins at the diagnosis of single death and during ultrasonographic follow up. In monochorionic pregnancies prenatal MRI, when performed, was negative for signs of brain damage in the surviving cotwins. Gestational age at delivery was not statistically different between monochorionic and dichorionic pregnancies (36 (range, 28.4–40.2) vs. 34.6 (range, 28.2–41.3) weeks) (p = 0.6) and the rate of early preterm delivery before 32 weeks was 23.8% (5/21) and independent from chorionicity (18.2% vs. 30%, p = 0.5). Neurodevelopmental follow up was available for 18/20 live born survivors (85%) and was normal in all but one twin; this case was born from a dichorionic pregnancy with a suspicion of congenital infection.

Conclusions

Our data confirmed a trend to a higher risk of perinatal mortality of cotwins in monochorionic twin pregnancies compared to dichorionic ones. In our experience prenatal ultrasound and MRI were useful to exclude cerebral lesions in utero and subsequent neurological sequelae in surviving monochorionic cotwins, even if definitive conclusions, especially on MRI, are limited by the small number of cases in our study.     

Congenital cytomegalovirus infection among twin pairs

Objective: The purpose of this study was to describe the fetal/neonatal cytomegalovirus (CMV) status according to chorionicity and outcome in twin pregnancies diagnosed with CMV.

Methods: An opportunistic diagnosis of CMV infection was performed in a tertiary referral center. All cases diagnosed in twin pregnancies (2006–2011) were included. Prenatal diagnosis was performed by CMV-DNA in the amniotic fluid (AF) of both fetuses only on the evidence of sonographic findings in either one or both twins. Neonatal screening was selectively assessed in symptomatic newborns, preterm, and infants born to HIV-infected mothers. Congenital infection was considered in the presence of CMV-DNA in AF, fetal tissues or newborn urine within the first 2 weeks of life, and symptomatic disease with clinical findings at birth or autopsy.

Results: A total of six twin pregnancies with congenital CMV infection were diagnosed, five dichorionic and one monochorionic diamniotic. Only one sibling was infected among dichorionic pregnancies, two diagnosed prenatally, and three after birth. In the monochorionic pregnancy, the diagnosis was performed prenatally and the two fetuses were infected and severely damaged.

Conclusions: Congenital CMV infection in twins might be related, among other factors, to chorionicity, and in DC twins a non-concordant infection can be expected.     

Application of QF-PCR for the prenatal assessment of discordant monozygotic twins for fetal sex

OBJECTIVE: To establish the utility of quantitative fluorescent polymerase chain reaction (QF-PCR) in order to determine the zygosity of multiple pregnancies, as well as to define the origin of the most frequent aneuploidies in amniotic fluid samples. METHODS: We describe the case of a monochorionic (MC) diamniotic (DA) pregnancy with phenotypically discordant twins (nuchal cystic hygroma and non-immune hydrops in twin A and no anomalies in twin B). QF-PCR was performed for rapid prenatal diagnosis in uncultured amniocytes and subsequently in cultured cells. Polymorphic markers for chromosomes X, Y, 13, 18 and 21 were used for determination of zygosity as well as sex chromosome aneuploidy. RESULTS: Twin A showed a Turner Syndrome (TS) mosaicism pattern by QF-PCR in uncultured amniocytes. The monozygotic origin of the pregnancy was determined. Interphase fluorescence in situ hybridization (I-FISH) in this sample showed a mosaicism X0/XY (83/17%). Cytogenetic analysis revealed a 45,X0 karyotype in twin A and a 46,XY karyotype in twin B. CONCLUSIONS: QF-PCR is a reliable tool for the determination of the zygosity independently of the chorionicity and the fetal sex in case of twin pregnancy. Testing both direct and cultured cells can provide useful results for genetic counselling in chromosomal mosaicisms.     

Selective feticide in dichorionic pregnancies with intracardiac blood aspiration: report of nine cases

In a twin pregnancy discordant for a chromosomal or structural abnormality, the family is faced with two alternatives: expectant management or selective feticide of the abnormal fetus. Methods of selective feticide differ according to chorionicity due to the presence of intertwin anastomoses in monochorionic placentas. Intracardiac KCl injection is commonly used in dichorionic pregnancies. Here, we describe an alternative method for selective feticide in dichorionic pregnancies by blood aspiration from the fetal heart until cardiac arrest.     

Prevalence of cranial scan abnormalities in preterm twins in relation to chorionicity and discordant birth weight

OBJECTIVE: The purpose of this study was to determine the incidence of ultrasonographically detected cerebral white matter lesions (WMLs) in preterm twins at birth in relation to chorionicity, discordant weight and twin-twin transfusion syndrome (TTTS). METHODS: In this retrospective study, perinatal, neonatal, and cranial scan data of 85 monochorionic (MC) and 94 dichorionic (DC) twin pregnancies (341 infants) delivered between 24 and 34 weeks of gestation were collected. Data were analysed according to chorionicity, discordant birth weight (>20%), single intrauterine death and TTTS. RESULTS: The cerebral WML was seen in 14% of preterm twins. Monochorionic infants had higher risks of WML than DC twin (odds ratio 7.1; 95% CI 3.28-15.8). In MC group, discordant weight (37%), TTTS (38%), single intrauterine death (67%) had higher incidence of cerebral WML than concordant weight infants (7%). Similarly, incidence of WML was higher in DC discordant compared with concordant weight infants (13% versus 2%; P < 0.05). CONCLUSION: Monochorionic infants had a seven-fold higher incidence of cerebral WML than DC infants. Discordant birth weight, TTTS and survivor of co-twin demise are an independent risk of cerebral white matter lesion.     

Perinatal outcomes of twin pregnancies discordant for major fetal anomalies

OBJECTIVE: The aim of this study was to determine perinatal outcomes of twin pregnancies discordant for a major fetal anomaly and to compare with twins without anomaly. METHODS: All twin pregnancies admitted or referred to the maternal-fetal unit were prospectively entered into a computer database. Chorionicity, fetal anomaly, mean gestational age at delivery, birth weight and perinatal survival rate were reviewed. MAIN OUTCOME MEASURES: Mean gestational age at delivery, birth weight and perinatal survival rate of twins with and without anomaly. RESULTS: There were 48 cases of monochorionic diamniotic (MCDA), 2 cases of monochorionic monoamniotic (MCMA) and 217 twins with dichorionic (DC) placentation. Out of 267 twin pregnancies, there were 17 (6.3%) twins with fetal anomaly. Twins discordant for a major fetal anomaly were diagnosed in 13 cases (4.8%). We observed 3 cases with MCDA and 10 cases with DC placentation and the incidence of discordance for a major fetal anomaly as 4.6% (10/217) in DC and 6.0% (3/50) in MC twin pregnancies. We identified 8 cases (62%) with craniospinal, 2 (15%) with gastrointestinal, 2 (15%) with urinary system, and 1 case (8%) with both craniospinal and gastrointestinal anomalies. There were significant differences between the normal co-twin of the major anomaly group (n = 13) and twins without anomaly group (n = 235) in mean gestational age at delivery (32 vs. 34 weeks; p = 0.029), mean birth weight (1,640 vs. 2,030 g; p = 0.022) and perinatal survival rate (69.2 vs. 91.1%; p = 0.018), respectively. CONCLUSION: The presence of a fetus with a major anomaly in a twin gestation increases the risk of preterm delivery, low birth weight and perinatal mortality of the normal co-twin.     

Strategies of antenatal diagnosis and management of pathologies related to multiple pregnancies

In the first trimester, ultrasound is crucial to screen for aneuploidies based on nuchal translucency, to diagnose major birth defects, and to establish chorionicity. Indeed prenatal diagnosis strategies as well as obstetrical management options are largely based on placental type. In the second trimester, maternal serum screening is not as effective, and ultrasound screening is more difficult in twins than in singletons. When a severe abnormality in found in one twin, selective termination of pregnancy may be considered. This technique is safe in dichorionic twins, but hazardous in monochorionic pregnancies. Selective termination in dichorionic twins is safer in the first trimester, underscoring the need for early prenatal diagnosis in twins. Monochorionic twin pregnancies carry specific risks, such as the twin-to-twin transfusion syndrome, which can be treated by endoscopic photo-coagulation of intertwin anastomoses in the severe early onset cases, or by amnioreduction in milder cases.     

Intertwin crown-rump length discordance in the prediction of fetal anomalies,fetal loss and adverse perinatal outcome

Objective: To assess the risks of fetal anomalies, fetal loss and adverse perinatal outcome in a cohort of first-trimester intertwin crown-rump length (CRL) discordant twins, stratified by chorionicity and the degree of CRL discordance.

Method: Four-hundred-and-seventy-one twin pregnancies were scanned during an 8-year period at 11–14 weeks, and those with an intertwin CRL discordance ≥10% were compared with concordant twins. Outcomes were also compared between monochorionic and dichorionic twins and between moderate (10–16%) and severe (>16%) discordance.

Results: Four-hundred-and-five twin pregnancies, 65 discordant and 340 concordant, were follow-up. Discordant twin pregnancies were at significant higher risk of chromosomal (OR?=?11.42; 95%?CI: 2.78–46.94) and structural anomalies (OR?=?5.91; 95%?CI: 2.25–15.54), spontaneous fetal loss (OR?=?4.23; 95%?CI: 1.79–10.01), birthweight discordance (OR?=?2.8; 95%?CI: 1.48–5.65) and small-for-gestational age (OR?=?3.48; 95% CI: 1.78–6.79). Similar differences (except for birthweight discordance) were observed among dichorionic twins. Among monochorionic, increased frequencies were only seen for structural anomalies, birthweight discordance and small newborns. Severe CRL discordance presented with higher rates of structural anomalies, stillbirth, birthweight discordance and small newborns.

Conclusion: Intertwin CRL discordance (≥10%) results in an increased risk of fetal anomalies and growth restriction that increases in severe CRL discordance (≥16%).     

Twin pregnancy outcome and chorionicity

BACKGROUND: Based on research of relatively poor quality, it is generally believed that dichorionic twins have lower perinatal mortality than monochorionic twins. We assessed the relationship between the pattern of perinatal loss in twin pregnancy and chorionicity. METHODS: A cohort study of 238 consecutive sets of twin pregnancies registered in our antenatal service over a 2-year period (1996-98) had chorionicity determined by ultrasound at 12-16 weeks' gestation. Follow up included scanning at 20 weeks' gestation for anomaly, and at 2-3-weekly intervals from 24 weeks' gestation onwards for growth and well being. Hazard ratios were computed for comparing the risk of death according to chorionicity. The perinatal loss patterns were analyzed according to gestational age at demize and that at delivery using survival analysis. RESULTS: Overall, 17/238 (7.1%) twin pregnancies suffered mortality: 14/190 (7.3%) amongst the dichorionic and 3/48 (6.2%) amongst the monochorionic twins. The hazard ratio for mortality was 0.89 (95% confidence interval 0.27-2.97, p = 0.85), considering gestational age at demize as the outcome. For gestational age at delivery as the outcome, the hazard ratio for mortality was 0.93 (95% confidence interval 0.27-3.15, p = 0.91). Survival analysis showed that amongst the dichorionic twins the hazard of death continued to rise throughout gestation. In contrast, the hazard of death for the monochorionic twins rose gradually to a maximum at 28 weeks' gestation and was then constant. CONCLUSION: Chorionicity did not affect the overall fetal loss rate amongst the twin pregnancies in our setting. There were differences in patterns of loss according to chorionicity, which require further investigation.