Manage twin and higher-order multiple pregnancies, including TTTS, sIUGR and single demise

In one line

Chorionicity — not zygosity — drives risk, surveillance and decision-making in multiples; every monochorionic pair shares a placenta and therefore an open vascular circuit, so the consultant's task is to detect the inter-twin transfusion syndromes (TTTS, TAPS, sIUGR) early and to anticipate that demise of one fetus acutely threatens the survivor's brain.

This chapter assumes the multiple pregnancy basics from the Intermediate course (zygosity, the lambda/T-sign, dichorionic surveillance) and spends its words on advanced monochorionic management and its unresolved controversies.

Why this matters in South Africa

The consultant problem in SA is not the diagnosis — it is the geography. Every monochorionic complication is a fetal-therapy problem, and fetoscopic laser is available at only a handful of tertiary units in the country. The clock that matters in TTTS (refer before the dividing membrane is too stuck for the fetoscope, ideally weeks 16–26) collides with district→regional→tertiary referral latency, transport, and bed availability. The judgement is therefore twofold: (1) recognise the monochorionic complication a fortnight earlier than a district sonographer would, because that fortnight is the difference between a laserable placenta and a salvage delivery; and (2) make the right system call — who can be surveilled regionally, who must be referred to the laser centre now, and who is beyond intervention and is simply being timed for delivery. Knowing the Quintero stages counts for little without knowing when to pick up the phone.

Pathophysiology — the shared circuit and its four failure modes

This assumes the placentation groundwork in the Intermediate chapter. Every monochorionic placenta is one organ with one vascular bed crossed by inter-twin anastomoses, and the four monochorionic syndromes are four distinct haemodynamic failures of that single shared circuit. The underlying mechanism is what separates the four — and it dictates the intervention.

• Anastomosis architecture is the master variable. Three connection types coexist on the vascular equator: artery-to-artery (AA), vein-to-vein (VV) and the deep, unidirectional artery-to-vein (AV) connections that run through a shared cotyledon. AV anastomoses are one-way pumps; net transfusion occurs when AV flow in one direction is not balanced by an AA anastomosis (which is bidirectional and pressure-equalising and therefore protective). The presence and calibre of a large AA anastomosis is the single best protector against TTTS and the reason sIUGR type II/III behaves as it does.
• TTTS = a volume/pressure imbalance. Net AV transfusion makes the donor hypovolaemic (oliguria → oligohydramnios → "stuck twin") and the recipient hypervolaemic (polyuria → polyhydramnios). The recipient's volume load drives the renin–angiotensin paradox: the donor's under-perfused kidney releases renin, vasoactive mediators cross to the recipient, and the recipient develops hypertension, myocardial hypertrophy and ultimately a cardiomyopathy with functional/structural right-ventricular outflow obstruction — which is why the sicker heart is in the bigger twin. This is a fluid-and-pressure disease, so the fluid discordance is the diagnostic signal.
• TAPS = a slow, isolated red-cell imbalance through tiny anastomoses. When transfusion occurs only through a few minuscule (<1 mm) AV anastomoses with no compensating AA, the transfer is too slow to shift volume (so liquor stays normal) but chronic enough to make the donor anaemic and the recipient polycythaemic. It is the haemoglobin-without-fluid disease, and it is the classic post-laser complication when a residual tiny anastomosis is left behind. Hence it is screened by MCA-PSV, not by fluid.
• sIUGR = unequal placental sharing. Unequal territorial division of the single placenta starves one twin. The small twin's umbilical-artery waveform — and crucially the behaviour of a large AA anastomosis — defines the Gratacós types and the mode of death (see Assessment). The mechanism is placental real-estate, not transfusion, which is why fluid is usually normal.
• TRAP (acardiac twin) = a reversed, parasitic circuit. Through a large AA (and usually VV) anastomosis present from early development, the "pump" twin perfuses an acardiac, structurally absent co-twin in reverse (deoxygenated blood enters the recipient via its umbilical artery). The pump twin is at risk of high-output cardiac failure and polyhydramnios, not anaemia. Occurs in roughly 1% of monochorionic pregnancies; the threat is to the structurally normal pump twin.

The clinical corollary: a monochorionic placenta cannot be managed as "twice a singleton." Anything you do to one twin (deliver it, let it die, transfuse it) is transmitted through the shared circuit to the other within minutes. This is the unifying reason single demise damages the survivor, reduction must occlude the cord rather than inject the fetus, and laser must be complete.

Assessment

• Chorionicity by 14 weeks is the single most important determination. It is reliable only in the first trimester: the lambda (twin-peak) sign = dichorionic, the T-sign = monochorionic. The SA SAJOG best-practice guideline requires it documented before 13⁺⁶ weeks; after this, membrane-count and fetal-sex inference are unreliable, and an undated "twin pregnancy" referred late should be treated as monochorionic until proven otherwise.
• Label the twins unambiguously (e.g. left/right, or upper/lower relative to a fixed landmark) and keep the mapping for the whole pregnancy — discordance is meaningless if you cannot say which fetus is which between scans.
• Monochorionic diamniotic (MCDA) surveillance is fortnightly from 16 weeks. At each scan: deepest vertical pool in both sacs, bladder visibility, EFW for discordance, umbilical-artery Doppler, and MCA peak systolic velocity (MCA-PSV) to screen for TAPS. Document the donor bladder and stomach.
• Read the pattern, not the single number. TTTS = polyhydramnios/oligohydramnios sequence (DVP >8 cm recipient before 20 wks, or >10 cm after; <2 cm donor) — amniotic-fluid discordance, not size discordance. TAPS = isolated haemoglobin discordance with normal fluid: MCA-PSV >1.5 MoM (anaemic donor) and <1.0 MoM (plethoric recipient), or a delta-MCA-PSV >0.5 MoM. sIUGR = EFW discordance ≥25% with a small twin <10th centile — a growth problem classified by the small twin's umbilical-artery Doppler.
• Higher-order multiples add the chorionicity permutations (a triplet can be DCTA, MCTA, etc.); each monochorionic component carries the monochorionic risks, and fetal reduction/selective-termination counselling becomes part of assessment.

Staging the syndromes — and where each system misleads

Beyond recognising TTTS, the task is staging it correctly, because the stage sets the intervention threshold and the prognosis.

• TTTS — Quintero I–V. Stage I: discordant fluid only, donor bladder still visible. Stage II: donor bladder no longer visible (the watershed — bladder absence means the donor is profoundly oliguric). Stage III: critically abnormal Dopplers in either twin (absent/reversed UA end-diastolic flow, reversed ductus venosus a-wave, or pulsatile umbilical vein). Stage IV: hydrops (usually the recipient, from cardiac failure). Stage V: demise of one or both. The system's well-known weakness is that it is not strictly sequential and not purely prognostic — a pregnancy can present at stage III without passing through II, and recipient cardiomyopathy (the strongest predictor of recipient death) is not captured at all. This is why some centres add a cardiovascular (CHOP-type) score; Quintero remains the common staging language, but recipient cardiac function and cervical length modify the plan independently of stage.
• TAPS — Slaghekke antenatal 5-stage. Stage 1: donor MCA-PSV >1.5 MoM and recipient <1.0 MoM, no other compromise. Stage 2: donor >1.7 MoM and recipient <0.8 MoM, no other compromise. Stage 3: stage 1 or 2 plus cardiac compromise of the donor. Stage 4: hydrops of the donor. Stage 5: demise of one or both preceded by TAPS. Note the asymmetry: in TAPS it is the donor (the anaemic twin) that decompensates and becomes hydropic — the mirror image of TTTS, where the hydropic twin is the recipient. The newer delta-MCA-PSV >0.5 MoM criterion (Tollenaar) improves antenatal detection over the absolute 1.5/1.0 cut-offs and is the modern screening threshold.
• sIUGR — Gratacós I/II/III by the small twin's UA Doppler. Type I: positive UA end-diastolic flow → benign, often deliverable near term. Type II: persistently absent/reversed end-diastolic flow → high risk of deterioration, plan delivery preterm. Type III: intermittently absent/reversed end-diastolic flow (the "iAREDF" pattern), caused by a large AA anastomosis intermittently dumping flow between twins. Type III is the trap: it can look stable for weeks and then cause sudden, unpredictable death of the small twin, with acute exsanguination of the larger twin into the dying small one — the same mechanism as single demise. Type III therefore needs tertiary co-management and a low threshold for steroids, not routine fortnightly scanning.

The differential — the discordances that are NOT a transfusion syndrome

The consultant error is to label any discordance "TTTS" and refer for laser. Map the discordance to its mechanism first:

• Size discordance with normal fluid and normal MCA-PSV = sIUGR, not TTTS. Laser is not the default; the question is timing by Gratacós type.
• Fluid discordance that is actually a structural cause — a recipient with isolated polyhydramnios may have an oesophageal atresia or a cardiac lesion; a "donor" with anhydramnios may have renal agenesis or PPROM of one sac. Always exclude an anomaly and ruptured membranes before calling TTTS.
• MCA-PSV discordance in a dichorionic pair is not TAPS (no shared anastomoses) — it is alloimmune or parvovirus anaemia of one twin, managed as a singleton anaemia.
• A discordant nuchal translucency / early growth lag in the first trimester predicts later TTTS/sIUGR but is not yet a syndrome — it flags the pregnancy for closer surveillance.
• Selective fetal anomaly or aneuploidy in one twin changes the entire counselling and the reduction technique (cord occlusion, never intracardiac KCl, in a monochorionic pair — KCl would cross to the normal twin).
• TRAP can masquerade as a "demised twin with persisting growth" — the giveaway is reversed perfusion (arterial-direction flow entering the mass) on Doppler.

Management

Immediate (the acute transfusion syndromes):

• TTTS untreated below 26 weeks carries ~80–90% perinatal loss; laser is first-line over amnioreduction or septostomy (Eurofetus). Refer before the membrane becomes too stuck for fetoscopy.
• For single intrauterine demise of a monochorionic co-twin, the survivor's hypotensive exsanguination into the dead twin has already occurred. Counsel the parents that ~1 in 4 survivors has neurological injury and ~15% die. Offer fetal MRI at 3–4 weeks (earlier brain injury is invisible) and survivor MCA-PSV (anaemia can be salvaged by intrauterine transfusion). Deliver for obstetric reasons or at term-appropriate gestation, not in panic.

Subtype-specific management in depth

TTTS — what the laser actually does, and the named techniques. Fetoscopic laser ablates the anastomoses on the vascular equator, converting one shared placenta into two functionally independent ones. The named techniques and how they differ:

• Selective laser — coagulate only the anastomoses you can identify as crossing the inter-twin membrane. Faster, but leaves the placenta with un-mapped residual connections → post-laser TAPS and recurrent TTTS.
• Sequential selective laser — occlude the AV anastomoses from donor to recipient first (stopping the net transfusion before reversing it), then the recipient-to-donor and finally the AA/VV. A refinement intended to improve donor survival.
• Solomon technique — after selective coagulation, draw the laser along the entire vascular equator to dichotomise the whole placental surface, deliberately catching the small anastomoses a selective approach misses. This is the modern default (see evidence): it cuts post-laser TAPS and recurrence dramatically at no survival cost.

TTTS — the stage I controversy, stated precisely. Stage I is the only genuinely unsettled threshold. A meaningful fraction of stage I cases are stable or regress, and a randomised trial plus meta-analysis show no significant survival advantage of routine early laser over expectant management in asymptomatic, stable stage I disease with a normal cervix (meta-analysis OR for at-least-one survivor ~1.40, 95% CI 0.26–7.43 — wide and non-significant; double-survival OR ~1.63, 0.74–3.62). But ~60% of stage I progresses, so the defensible plan is expectant management with twice-weekly surveillance at a centre that can laser within hours, with laser triggered by progression to stage II, maternal symptoms, or a short cervix. Do not state "always laser stage I"; do not state "never refer stage I."

TAPS — four options matched to stage and gestation. There is no single right answer; the consultant tailors:

• Expectant + intensified surveillance — stage 1, remote from term.
• Intrauterine transfusion of the donor (± partial exchange transfusion of the polycythaemic recipient) — buys time when laser is not feasible or the membrane precludes fetoscopy; treats the anaemia but not the underlying anastomosis, so it recurs.
• Fetoscopic laser — the only definitive treatment, but technically hard because the tiny anastomoses are difficult to see; reserved for earlier gestations at expert centres.
• Delivery — near term, or for higher stage with fetal compromise. The recipient's polycythaemia/hyperviscosity needs neonatal awareness at birth.

sIUGR — timing by type, not by size. Type I: serial growth/Doppler, aim for ~34–36 weeks. Type II: expect deterioration; give steroids and plan delivery typically in the early-30s of weeks, individualised on the small twin's ductus venosus and CTG. Type III: the danger is sudden death of the small twin taking the normal twin's brain with it, so the consultant either accepts intensive surveillance with steroids on board and a pre-agreed delivery gestation, or — when the small twin is pre-viable or moribund and threatening the normal co-twin — discusses selective reduction by cord occlusion to protect the survivor.

Selective reduction / selective feticide in a monochorionic pair. Because anything injected crosses the shared circuit, intracardiac KCl is contraindicated — feticide must mechanically occlude the cord/vessels: bipolar cord coagulation, radiofrequency ablation (RFA), or interstitial laser. Survival of the retained co-twin after these procedures is in the order of ~75–90% depending on technique and indication, and the procedure itself carries a risk of PPROM and preterm birth. Indications include a discordant lethal anomaly, severe early sIUGR threatening the normal twin, and TRAP.

TRAP sequence. The pump twin is the patient. When the acardiac mass is large relative to the pump twin or the pump twin shows cardiac strain/polyhydramnios, intrafetal ablation of the acardiac twin's feeding vessels (RFA, usually ≥16 weeks) protects the pump twin, with pump-twin survival reported above 90% in good series. Small acardiac masses with a well twin may be surveilled.

Ongoing surveillance and corticosteroids: continue chorionicity-appropriate scans; give a single course of antenatal corticosteroids when preterm birth (spontaneous or planned) is anticipated, and magnesium sulphate for neuroprotection if delivering <32 weeks — the preterm labour and PPROM principles apply, but routine progesterone, cerclage and pessary do not reduce twin preterm birth and should not be offered for multiplicity alone. See preterm-labour-and-pprom and fetal-growth-restriction.

Long-term / timing & mode: plan delivery by chorionicity — uncomplicated DCDA by 37⁺⁰ weeks, uncomplicated MCDA by 36⁺⁰ (after steroids), and MCMA by 32–34 weeks via caesarean (cord entanglement). Vaginal birth is appropriate when twin 1 is cephalic; the second-twin skill set (internal podalic version, breech extraction, ultrasound on the labour ward) must be available, which in the SA district→regional→tertiary system means MCDA and any complicated multiple is a regional/tertiary delivery, not a district one.

The monoamniotic (MCMA) problem — cord entanglement and where to surveil

MCMA twins (no dividing membrane) carry a unique, unpredictable stillbirth risk from cord entanglement that is not reliably warned by routine Doppler — entanglement is near-universal and most do not occlude, so the death is sudden. The consultant decisions:

• Where to surveil. The MONOMONO comparison found no statistically significant difference in fetal death between inpatient continuous-monitoring protocols (from ~26–28 weeks) and structured outpatient surveillance (from ~30 weeks) — at a large cost in maternal hospital days for inpatient care. So inpatient admission is defensible, not mandatory; the honest SA answer balances bed availability, distance from the hospital, and the family's ability to return urgently against a real but modest absolute risk.
• When to deliver. The prospective risk of intrauterine death persists each week into the early 30s (reported around 5–7% per week at 32–34 weeks), and falls sharply once delivered; the consensus is caesarean at 32–34 weeks after steroids, not expectant management to term.
• How to deliver. Caesarean is standard precisely because cord entanglement makes a vaginal second-twin manoeuvre dangerous.

The evidence & the controversy

Laser, and which laser. The Eurofetus RCT established fetoscopic laser as superior to serial amnioreduction for severe TTTS, with higher survival and far less neurological morbidity. The unresolved question was residual anastomoses causing post-laser TAPS and TTTS recurrence; the Solomon trial answered it — coagulating the entire vascular equator cut TAPS from 16% to 3% and recurrence from 7% to 1%, and is now the default technique. The live controversy is Quintero stage I: many are stable or regress, and whether to laser, amnioreduce or observe stage I remains genuinely unsettled — the randomised and pooled data show no clear survival gain from routine early laser in asymptomatic stable stage I, so the defensible approach is individualised expectant management with close surveillance at a fetal-therapy centre and laser triggered by progression, symptoms or a short cervix.

Timing of delivery is a long-standing controversy. The Cheong-See BMJ meta-analysis showed the stillbirth-versus-neonatal-death curves cross at 37 weeks for dichorionic twins (delaying to 38 weeks adds ~8.8 perinatal deaths/1000). For monochorionic twins the data were less certain, supporting 36 weeks as a pragmatic compromise rather than a hard rule — NICE and the SA guideline both land here. Do not over-read the monochorionic figure as proven.

Mode of delivery. The Twin Birth Study (Barrett) randomised cephalic-leading twins 32⁺⁰–38⁺⁶ weeks and found no difference in death or serious morbidity between planned caesarean and planned vaginal birth (2.2% vs 1.9%). This underpins offering vaginal birth when twin 1 is cephalic and a clinician skilled in second-twin delivery is present — directly relevant to SA, where blanket caesarean for twins wastes scarce theatre capacity without benefit. The caveat: the trial excluded monoamniotic twins and required the skills the result depends on.

Quantifying the single-demise risk. The Hillman meta-analysis supplies the numbers: after single fetal death, the monochorionic co-twin dies in ~15% and has neurodevelopmental impairment in ~26%, versus ~3% and ~2% for dichorionic — roughly a five-fold excess (OR ~4.8 for impairment). This is the arithmetic that justifies not delivering reflexively (the injury is already done at the moment of demise; delivery only adds prematurity to the survivors who are not injured) and that frames the delayed-interval MRI and survivor-MCA-PSV plan.

Landmark trials & key evidence

Worked viva — how to structure the answer

Take a stem such as "A 24-year-old, MCDA twins at 21 weeks, referred with a 'big twin and a small twin'; the recipient sac DVP is 11 cm, the donor sac 1 cm, donor bladder not seen."

1. Classify it precisely — "This is the poly/oligo fluid sequence with an empty donor bladder, so this is TTTS, Quintero stage II — not sIUGR or TAPS; the discordance is in fluid, which fixes the diagnosis."
2. State the threat and the window — "Untreated below 26 weeks this carries ~80–90% perinatal loss; the treatment is fetoscopic laser and it is time-critical because a stuck membrane will preclude fetoscopy."
3. Act on the system — "I would refer urgently to the nearest fetal-therapy centre that performs laser, today, with the scan images, and counsel the parents; in SA that means arranging tertiary transfer, not local management."
4. Name the technique and why — "Solomon whole-equator laser is preferred over selective laser because it reduces post-laser TAPS and recurrence, and over amnioreduction because Eurofetus showed better neurologically-intact survival."
5. Plan the aftercare — "Post-laser surveillance for TAPS by MCA-PSV and for recurrence; steroids and magnesium neuroprotection if early delivery looms; if one twin demises, I do not deliver reflexively — survivor MCA-PSV and delayed-interval MRI, because the insult is at the moment of death."
6. Close honestly on uncertainty — "Had this been stable, asymptomatic stage I with a normal cervix, I would have offered close surveillance rather than reflex laser, because the randomised data show no clear survival gain from routine early laser there."

Exam traps & red flags

• Confusing the three syndromes. TTTS = fluid discordance; TAPS = haemoglobin (MCA-PSV) discordance with normal fluid; sIUGR = size discordance. Naming the wrong one mis-directs the entire plan.
• Mixing up which twin decompensates. In TTTS the hydropic/cardiomyopathic twin is the recipient; in TAPS the hydropic twin is the donor (the anaemic one). Getting it the wrong way round means the mechanism has been misunderstood.
• Delivering a single monochorionic demise as an emergency. The brain injury already happened at demise; immediate delivery only adds iatrogenic prematurity. The correct moves are survivor MCA-PSV, delayed-interval MRI and counselling.
• Treating a late-booked twin as dichorionic. If chorionicity was never established in the first trimester, manage as monochorionic — the higher-surveillance, safer error.
• Offering progesterone, cerclage or pessary "because it's twins." None reduces twin preterm birth; reserve them for the individual indications covered under preterm labour.
• Missing sIUGR type III's sudden death. Intermittent absent/reversed UA flow can kill the small twin without warning and damage the survivor — this needs tertiary co-management, not routine fortnightly scans.
• Using intracardiac KCl for selective reduction in a monochorionic pair. It crosses the shared circuit and kills the normal twin — reduction must mechanically occlude the cord (bipolar/RFA/laser).
• Forgetting the recipient's heart. TTTS recipients develop cardiomyopathy and can have functional/structural RVOT obstruction — fetal echo and paediatric-cardiology awareness matter.
• Calling any fluid excess "TTTS." Exclude a structural anomaly (recipient cardiac/oesophageal lesion) and PPROM of one sac before labelling and referring for laser.
• MCMA twins managed like MCDA — no dividing membrane means cord entanglement; deliver by caesarean at 32–34 weeks with intensified surveillance.

Evidence anchors

• Senat MV et al. Endoscopic laser surgery vs serial amnioreduction for severe TTTS. N Engl J Med 2004 (Eurofetus)
• Slaghekke F et al. Fetoscopic laser coagulation of the vascular equator vs selective coagulation for TTTS (Solomon). Lancet 2014;383:2144–51
• Slaghekke F et al. Twin anemia-polycythemia sequence: diagnostic criteria, classification, perinatal management and outcome. Fetal Diagn Ther 2010
• Hillman SC, Morris RK, Kilby MD. Co-twin prognosis after single fetal death: systematic review & meta-analysis. Obstet Gynecol 2011
• Cheong-See F et al. Prospective risk of stillbirth and neonatal complications in twin pregnancies. BMJ 2016;354:i4353
• Barrett JFR et al. A randomized trial of planned cesarean or vaginal delivery for twin pregnancy. N Engl J Med 2013;369:1295–1305
• Gratacós E et al. A classification system for selective FGR in monochorionic pregnancies. Ultrasound Obstet Gynecol 2007
• Quintero RA et al. Staging of twin-twin transfusion syndrome. J Perinatol 1999;19:550–5
• Tollenaar LSA et al. Improved prediction of TAPS by delta MCA-PSV: new antenatal classification. Ultrasound Obstet Gynecol 2019
• Fetoscopic laser photocoagulation vs expectant management for stage I TTTS: systematic review & meta-analysis. Prenat Diagn 2023
• D'Antonio F et al. (MONOMONO) Inpatient vs outpatient management and timing of delivery of uncomplicated monochorionic monoamniotic twin pregnancy. Ultrasound Obstet Gynecol 2019
• NICE NG137. Twin and triplet pregnancy (updated 2024)
• RCOG Green-top Guideline No. 51. Management of monochorionic twin pregnancy (2024 partial update)
• South Africa SAJOG best-practice guideline: prenatal care for twin pregnancies (chorionicity by 13⁺⁶ weeks; MFM referral) — plain text, journal-stable URL varies.
• SA NDoH Guidelines for Maternity Care in South Africa (referral level for multiples; corticosteroids; MgSO₄ neuroprotection) — plain text.