Twin pregnancy is a pregnancy in which two embryos or fetuses develop simultaneously. Twins may arise from two separately fertilized eggs or from a single fertilized egg that gives rise to two embryos. These origins determine their genetic relationship, but the most important distinction for obstetric care is often whether they share a placenta and whether they occupy separate amniotic sacs.
Compared with singleton pregnancy, twin pregnancy places greater demands on maternal physiology and carries higher risks of premature birth, impaired fetal growth, hypertensive disorders, and complications around delivery. Twins sharing a placenta also face distinctive risks caused by vascular connections between their circulations. Many twin pregnancies nevertheless result in two healthy infants, with outcomes strongly influenced by gestational age at birth, placental arrangement, associated abnormalities, and access to specialized care.
Embryological origins and placental arrangements
Dizygotic twins, also called fraternal twins, develop when two eggs are fertilized by two sperm. They are genetically comparable to other full siblings and may have the same or different chromosomal sex. Each embryo normally develops its own chorion, the outer fetal membrane associated with placental formation, and its own amniotic sac. Their placentas may remain separate or become fused into a single-looking mass. Consequently, the appearance of one placental mass does not by itself establish that the twins share a chorion.
Monozygotic twins, commonly called identical twins, originate from one fertilized egg. They usually have the same chromosomal sex and share nearly all their inherited genetic sequence, although post-fertilization mutations, chromosomal mosaicism, and differences in gene regulation can produce biological differences. Monozygotic twins may have separate or shared placental structures. Chorionicity describes the number of chorions; amnionicity describes the number of amniotic sacs. These classifications are related to, but distinct from, zygosity.
Three principal arrangements occur. Dichorionic diamniotic twins have two chorions and two amniotic sacs. Monochorionic diamniotic twins share a chorion and placenta but occupy separate sacs. Monochorionic monoamniotic twins share both the chorion and amniotic cavity. Roughly two-thirds to three-quarters of monozygotic twin pairs are monochorionic, whereas monoamniotic twins represent only a small minority. Almost all monochorionic twins are monozygotic, although rare exceptions have been documented, particularly after fertility treatment.
The traditional embryological model associates earlier separation of a single embryo with dichorionic twins, later separation with monochorionic diamniotic twins, and still later separation with monoamniotic twins. Commonly cited intervals are approximately the first three days after fertilization, days four to eight, and days eight to thirteen, respectively. These intervals are a teaching model rather than a directly observed timetable applicable to every pregnancy. Human twinning mechanisms are not completely resolved, and some observations from assisted reproduction do not fit a simple splitting model.
Conjoined twins are a rare form of monozygotic twinning in which the fetuses remain physically connected. They are generally monochorionic and monoamniotic, but the location and extent of shared organs vary considerably. For routine twin pregnancy assessment, the distinction between separate and shared placentas remains more immediately informative than the popular distinction between “identical” and “fraternal.”
Frequency and determinants
Twin frequency can be expressed in several ways: twin pregnancies per total pregnancies, twin deliveries per total deliveries, or twin-born infants per total live births. These measures are not interchangeable. One twin delivery contributes two infants to a birth-based numerator, while early losses may prevent a pregnancy initially containing two embryos from appearing in delivery statistics. Comparisons also depend on whether registries include stillbirths and on the gestational threshold used to define a birth.
A global analysis published in 2021 estimated that the twinning rate increased from approximately 9.1 twin deliveries per 1,000 deliveries in 1980–1985 to 12.0 per 1,000 in 2010–2015. The later period corresponded to about 1.6 million twin pairs born annually. Rates vary substantially between populations. Spontaneous dizygotic twinning is particularly frequent in parts of sub-Saharan Africa and historically less frequent in East Asia. Monozygotic twinning is more geographically consistent, occurring at approximately three to four pairs per 1,000 deliveries.
Dizygotic twinning is associated with multiple ovulation. Its frequency increases with maternal age through much of the reproductive period and is associated with previous births and a maternal family history of fraternal twinning. These relationships reflect differences in ovarian function and inherited susceptibility to releasing more than one egg. They do not imply that all familial or age-associated twin pregnancies are dizygotic, nor that a twin pregnancy necessarily indicates a hereditary trait.
Assisted reproductive technology and ovulation-inducing treatment have substantially changed twin epidemiology. Transferring more than one embryo during in vitro fertilization can produce dizygotic twins if two embryos implant. Ovarian stimulation can also lead to several eggs being available for fertilization. Monozygotic twinning occurs after assisted reproduction as well, and its frequency is somewhat higher than after spontaneous conception, although the contribution of individual laboratory procedures remains incompletely understood.
The rise in twin births in many higher-income countries during the late twentieth century reflected both fertility treatment and later childbearing. Subsequently, greater use of single-embryo transfer reduced treatment-associated multiple births in several settings. National trends therefore differ: a global increase over several decades does not mean rates are continuing to rise everywhere. Differences in fertility-service access, embryo-transfer practice, population age structure, and birth registration all affect measured frequency.
Maternal physiological adaptation and complications
Twin pregnancy amplifies many normal physiological adaptations of pregnancy. Maternal plasma volume, cardiac output, and oxygen consumption increase to support a larger combined fetal and placental mass. Expansion of plasma volume can exceed the increase in red-cell mass, producing more pronounced dilution of blood hemoglobin. Requirements for iron and other nutrients are also greater. These changes help explain why iron-deficiency anemia is more common in twin than singleton pregnancy, although a low hemoglobin concentration requires interpretation in its clinical and laboratory context.
Greater uterine size can intensify mechanical symptoms such as reflux, breathlessness, pelvic pressure, and back discomfort. Concentrations of some placental hormones are higher on average, and nausea and vomiting may be more prominent. However, symptom severity varies widely and cannot diagnose twins. Maternal weight gain also reflects multiple components—fetuses, placentas, fluid, increased blood volume, and maternal tissues—rather than fetal size alone.
Twin pregnancy is an established risk factor for preeclampsia, a pregnancy-associated disorder involving hypertension and maternal organ dysfunction, sometimes accompanied by placental dysfunction. Risk is commonly estimated at roughly two to three times that of singleton pregnancy, although estimates vary by population and accompanying risk factors. Greater placental mass and altered release of placental signaling molecules are among the mechanisms investigated. Maternal blood pressure, urine findings, symptoms, and laboratory measurements are therefore important components of surveillance; fetal wellbeing alone does not exclude maternal disease.
The association with gestational diabetes is less uniform across studies than that with preeclampsia. Maternal age, body composition, ethnicity, and fertility treatment can influence comparisons. Twin pregnancy also adds to the increased risk of thrombosis associated with pregnancy, particularly when accompanied by hospitalization, immobility, operative delivery, or other risk factors. Observed complications often result from several interacting exposures rather than twin gestation alone.
After delivery, postpartum hemorrhage is more likely because an overdistended uterus may contract inadequately, a condition termed uterine atony. Larger placental implantation areas, operative birth, and associated disorders may also contribute. Maternal complications are therefore relevant throughout pregnancy, labor, and the postpartum period. Assessment includes both the direct effects of carrying two fetuses and pre-existing conditions, such as chronic hypertension or cardiac disease, that can alter tolerance of the additional physiological demands.
Diagnosis, dating, and prenatal screening
Ultrasonography establishes the number of embryos or fetuses and evaluates their location and viability. Increased uterine size, pronounced pregnancy symptoms, or elevated biochemical markers may raise suspicion but are not diagnostic. Early imaging can also identify a “vanishing twin,” in which one embryo stops developing and is subsequently resorbed or becomes inconspicuous. This helps explain why the frequency of twin conception exceeds the frequency of twin birth.
Determining chorionicity is most reliable in the first trimester, ideally before 14 weeks. In dichorionic pregnancy, a wedge of chorionic tissue may extend between the layers of the intertwin membrane, producing the lambda or twin-peak sign. A thin membrane meeting the placenta at approximately a right angle, the T sign, supports monochorionic diamniotic pregnancy. The assessment also considers membrane thickness, placental masses, and other findings. Apparent absence of a dividing membrane requires careful evaluation before classifying a pregnancy as monoamniotic.
Accurate gestational dating matters because growth assessment, interpretation of screening tests, and decisions about birth all depend on gestational age. Early crown–rump-length measurements are central to dating spontaneous twin pregnancies; established protocols commonly use the larger twin when a discrepancy is present, avoiding underestimation caused by early growth restriction. Pregnancies conceived through assisted reproduction can be dated using the known treatment timeline. Consistent identification of each fetus—by location, placental attachment, and other distinguishing features—allows findings to be followed across examinations.
Screening for chromosomal conditions is more complex than in singleton pregnancy. Cell-free DNA screening analyzes placentally derived DNA fragments circulating in maternal blood. It performs well for Down syndrome in twin pregnancy, but evidence and performance are less extensive for some other chromosomal conditions. Unequal DNA contributions from two placentas, a low fetal fraction, and a vanished twin can complicate interpretation. A positive screening result does not necessarily identify which fetus is affected and is not equivalent to a diagnostic result.
Diagnostic testing may involve chorionic villus sampling or amniocentesis, with sampling strategies determined by placental and sac arrangement. Accurate mapping is necessary to attribute a result to the correct fetus and avoid inadvertently sampling the same compartment twice. Detailed anatomical ultrasound evaluates each fetus separately because shared genetic origin does not guarantee identical anatomy. Genetic counseling addresses the distinction between screening and diagnosis, test limitations, and the implications of concordant or discordant findings.
Antenatal surveillance and fetal growth
Antenatal surveillance is organized primarily around chorionicity, gestational age, maternal health, and evolving fetal findings. Dichorionic twins have largely separate placental circulations, while monochorionic twins require additional assessment for shared-circulation complications. Surveillance therefore involves more than measuring two fetal heart rates: it evaluates anatomy, growth trajectories, amniotic fluid, placental function, and, where indicated, blood-flow patterns using Doppler ultrasound.
Established protocols commonly assess uncomplicated dichorionic twins at intervals of about four weeks during the latter part of pregnancy. Monochorionic diamniotic twins are generally examined at least every two weeks from approximately 16 weeks, primarily to detect twin-to-twin transfusion and related complications. Exact schedules differ between health systems and change when abnormalities appear. Monoamniotic pregnancies require a distinct surveillance and delivery framework because the absence of a dividing membrane introduces additional cord-related risks.
Fetal growth restriction describes failure to achieve expected growth potential and is not synonymous with being small. Some small fetuses have normal placental function, whereas a fetus above a conventional size threshold may nevertheless show a concerning decline in growth or abnormal Doppler findings. Estimated fetal weight is calculated from ultrasound measurements, including head size, abdominal circumference, and femur length. Measurement error is clinically relevant, especially when differences between fetuses are modest.
Growth discordance is often expressed as
where represents estimated fetal weight. A discordance of 20–25% commonly prompts increased attention, but no single percentage determines diagnosis or timing of birth. Interpretation also incorporates the smaller twin’s weight percentile, interval growth, amniotic fluid, gestational age, and umbilical-artery Doppler findings. In monochorionic twins, unequal allocation of placental territory and the configuration of vascular connections are especially important.
Whether twin-specific or singleton growth charts best distinguish physiological smallness from pathological restriction remains an active research question. Twins tend to grow differently from singletons later in pregnancy, so singleton charts classify more twins as small. Conversely, a reference based on observed twin size can potentially normalize some growth impairment. Clinical assessment therefore integrates growth charts with physiological evidence rather than treating a percentile as a complete measure of health. Maternal blood pressure and laboratory assessment continue alongside fetal surveillance because maternal and placental complications can evolve together.
Complications of a shared placenta or amniotic sac
Most monochorionic placentas contain vascular connections between the twins’ circulations. These connections can be artery-to-artery, vein-to-vein, or artery-to-vein through a placental cotyledon. Their number, direction, and resistance determine whether exchange remains balanced or produces disease. A shared placenta is therefore not simply a common source of nutrients: it can create direct hemodynamic dependence between the fetuses.
Twin-to-twin transfusion syndrome develops in approximately 10–15% of monochorionic diamniotic pregnancies. An imbalance of blood flow causes the donor twin to develop reduced circulating volume and low urine production, producing oligohydramnios, while the recipient develops increased urine output and polyhydramnios. The recipient can develop cardiac dysfunction, and the donor can become severely compromised. Diagnosis is based principally on the characteristic amniotic-fluid imbalance, with bladder appearance, Doppler findings, hydrops, and fetal death contributing to severity staging. Growth discordance is not required.
Fetoscopic laser treatment targets the placental vessels connecting the twins. By coagulating these anastomoses, it seeks to separate the circulations rather than merely relieve excess fluid. It is an established treatment for significant disease in the mid-trimester, supported by randomized evidence published in 2004 and subsequent studies. Amniotic-fluid drainage can relieve uterine distension but does not remove the underlying vascular connections. Treatment outcomes depend on disease severity, gestational age, technical factors, and subsequent complications.
Twin anemia–polycythemia sequence involves slow transfusion through small placental connections, resulting in anemia in one twin and an abnormally high red-cell concentration in the other, usually without the marked fluid imbalance of classic transfusion syndrome. It may occur spontaneously or after laser treatment. Doppler measurement of middle cerebral artery peak systolic velocity helps identify the contrasting blood states. A different rare disorder, twin reversed arterial perfusion sequence, occurs when a severely malformed twin lacking a functional heart receives retrograde arterial blood from a structurally more normal “pump” twin, potentially causing cardiac overload.
Death of one monochorionic twin can cause an abrupt transfer of blood from the survivor into the low-pressure circulation of the deceased twin, leading to hypotension, death, or ischemic brain injury in the survivor. This mechanism differs from the generally lower immediate circulatory risk after the death of a dichorionic co-twin. In monoamniotic pregnancy, the fetuses additionally share an unrestricted space in which umbilical cords can become entangled or compressed. Cord entanglement is common, but its presence alone does not predict the timing or occurrence of an acute adverse event.
Preterm birth and pregnancy prolongation
Preterm birth, defined as birth before 37 completed weeks, is the most frequent major complication of twin pregnancy. In many contemporary populations, more than half of twins are born preterm, and roughly one in ten are born before 32 weeks. Exact proportions vary with the population, chorionicity, and obstetric practice. Prematurity may follow spontaneous labor or membrane rupture, or it may result from medically indicated delivery for maternal disease, fetal compromise, or a placental complication.
Several mechanisms contribute. Increased uterine distension can promote contractile signaling and cervical remodeling. Placental endocrine activity, inflammatory pathways, and membrane stress may also be involved. The greater frequency of preeclampsia and growth restriction increases indicated preterm delivery. Consequently, a treatment that acts on only one pathway may have limited effects across all twin pregnancies, and results from singleton pregnancies cannot automatically be transferred to twins.
Transvaginal measurement of cervical length helps estimate the probability of spontaneous preterm birth. A shorter cervix in the mid-trimester is associated with higher risk, but prediction is imperfect: many affected pregnancies do not deliver imminently, and some early births occur without a markedly shortened cervix. The effectiveness of preventive interventions has been difficult to establish because studies differ in cervical-length thresholds, gestational age at treatment, and inclusion of asymptomatic women versus those with cervical dilation.
Routine bed rest has not demonstrated a preventive benefit and can introduce harms associated with immobility. Evidence does not support routine progesterone treatment for all unselected twin pregnancies. For twins with a short cervix, evidence and recommendations differ between professional bodies: some guidance supports vaginal progesterone, while other guidance restricts its use or calls for further trials. Cervical cerclage is likewise not a routine intervention for all twins, although selected situations, particularly cervical dilation, are assessed separately. These distinctions illustrate why “twin pregnancy” alone is an insufficient treatment category.
When early birth appears likely, hospital management may include corticosteroids to accelerate fetal lung maturation and magnesium sulfate for fetal neuroprotection at early gestations, according to the clinical situation and applicable protocols. Tocolytic drugs can sometimes delay labor briefly, principally to permit other interventions or transfer, but do not reliably prevent preterm birth over the remainder of pregnancy. Trials of these measures have not always included enough twin pregnancies to estimate twin-specific effects precisely, so some practice incorporates evidence from broader preterm-birth populations.
Timing and mode of birth
Planning birth requires balancing the risks of continuing pregnancy against those of earlier delivery. For twins, this balance usually favors birth earlier than for uncomplicated singletons. Chorionicity is a major determinant: monochorionic pregnancies carry shared-placental risks that are absent in most dichorionic pregnancies, while monoamniotic pregnancies add the possibility of acute cord compromise. Maternal disease, fetal growth, Doppler abnormalities, membrane rupture, and previous obstetric history can all alter the timing.
Professional recommendations differ somewhat. For example, United Kingdom guidance commonly offers planned birth at 37 weeks for uncomplicated dichorionic diamniotic twins and at 36 weeks for uncomplicated monochorionic diamniotic twins. For uncomplicated monochorionic monoamniotic twins, planned birth is usually substantially earlier, around 32–34 weeks. Other national guidance uses somewhat different windows. These are population-based frameworks for otherwise uncomplicated pregnancies, not fixed dates applicable regardless of clinical findings.
Twin pregnancy does not invariably require cesarean delivery. Important considerations include the presentation of the first twin, gestational age, estimated weights, size discordance, fetal condition, previous uterine surgery, and the availability of clinicians experienced in twin vaginal birth. The presentation of the second twin can change after the first is born. This creates a need for immediate reassessment and access to ultrasound, obstetric intervention, anesthesia, and newborn resuscitation.
The international Twin Birth Study, published in 2013, was a randomized controlled trial involving pregnancies between 32 weeks and 38 weeks 6 days in which the first twin was head-down. Under its eligibility criteria and with experienced obstetric care, planned cesarean delivery did not significantly reduce fetal or neonatal death or serious neonatal morbidity compared with planned vaginal delivery. The result does not apply automatically to pregnancies with a non-cephalic first twin, monoamniotic twins, or other situations outside the trial criteria.
During vaginal birth, the two fetal heart rates must be distinguished reliably. After the first birth, assessment of the second twin includes presentation, heart rate, membrane status, and bleeding. Depending on the circumstances, the second twin may be born head-first, delivered by breech extraction, or require cesarean delivery. A combined vaginal and cesarean birth is possible, though uncommon. The increased risk of uterine atony also makes preparation for postpartum bleeding an important component of delivery care, regardless of the planned route.
Neonatal outcomes, family effects, and research
The principal determinants of newborn outcome are gestational age, birth weight, congenital abnormalities, and complications arising before or during birth. Premature twins may require neonatal intensive care for respiratory support, temperature control, feeding, and management of metabolic instability. Respiratory distress, infection, intestinal disease, and intracranial injury become more frequent with increasing prematurity. A twin born near term with appropriate growth has a markedly different risk profile from a very preterm twin affected by placental disease.
Low birth weight, conventionally defined as less than 2,500 grams, is common among twins, but it combines two different processes: shortened gestation and restricted growth. These have overlapping yet distinct implications. Birth-weight discordance may persist after delivery, and each infant requires individual assessment rather than an assumption that both have identical needs. Shared genetic origin also does not guarantee identical adaptation to birth, feeding ability, or subsequent development.
Twin-born children have higher population-level rates of some developmental impairments than singleton-born children, largely associated with prematurity, growth restriction, and antenatal complications. The excess risk is not uniform across all twins. Severe shared-placental disorders and injury following the death of a monochorionic co-twin constitute additional pathways to neurological disability. Follow-up studies therefore need to account for gestational age and specific complications rather than attributing outcomes to twinning itself.
Care of two infants can impose substantial practical and financial demands, especially after prolonged hospitalization or when infants are discharged at different times. Feeding may involve direct breastfeeding, expressed milk, formula, or combinations according to clinical circumstances and family choices. Maternal recovery, sleep disruption, and the demands of simultaneous care can affect parental mental health. The relationship between multiple birth and psychological outcomes is influenced by social support, infant illness, household resources, and previous mental-health conditions.
Twin pregnancies also create distinctive research challenges. Outcomes for the two infants are correlated, so analyses that treat them as independent observations can underestimate statistical uncertainty. Studies must distinguish pregnancy-level outcomes, such as preeclampsia, from infant-level outcomes, such as respiratory morbidity, and specify how outcomes affecting only one twin are counted. Rare monochorionic complications often require multicenter collaboration to assemble adequate study populations. Current investigations include more precise growth standards, prediction of preterm birth, refinement of fetal therapy, and longer-term outcomes after placental vascular disease. Classical twin research on heredity is a related but separate field: comparisons of monozygotic and dizygotic twins depend on assumptions about shared environments, including prenatal conditions that can differ substantially by chorionicity.