Diagnose and manage complicated and non-tubal ectopic pregnancy, including caesarean-scar, cervical, interstitial, ovarian and advanced abdominal pregnancy

In one line

A non-tubal ectopic implants where there is no muscle to contain it and no easy access to remove it, so it bleeds catastrophically and late — the consultant task is to locate it precisely before treating, match the treatment to the site, and never reach for the tubal reflex (systemic single-dose methotrexate or a simple salpingectomy) on a caesarean-scar, cervical, interstitial or abdominal pregnancy, where that reflex either fails or kills.

This chapter assumes the tubal groundwork — the discriminatory zone, the serial-hCG logic, single-dose methotrexate, salpingectomy versus salpingotomy. It covers the implantation sites that behave differently from the tube, where the literature is thin, where the guidelines diverge or fall silent, and where the wrong first move is irreversible.

Mechanism & pathophysiology

The danger of every non-tubal site follows from why the tube itself is dangerous. The tube has a thin muscular wall and no decidua, so trophoblast invades the wall directly, erodes vessels, and ruptures into the peritoneum once the conceptus outgrows the lumen. Every non-tubal site is a variation on that theme — trophoblast invading tissue that was never built to hold a pregnancy or to be operated on safely — and the danger of each site is predicted by two things: how vascular the bed is, and how little normal myometrium lies between the trophoblast and a free peritoneal or vaginal surface.

The caesarean-scar pregnancy is the one that has changed the specialty, because its incidence rises in lockstep with the caesarean rate. A caesarean leaves a hysterotomy scar that often heals with a wedge-shaped defect — the niche (isthmocele) — a pouch of deficient myometrium on the anterior lower segment. A blastocyst can implant into that niche or onto the scar, and because the niche has little or no muscle beneath it, invading trophoblast reaches the bladder and the uterine vessels early. The central concept follows: caesarean-scar pregnancy and placenta accreta spectrum are the same disease at different gestations. A scar pregnancy that is left to continue does not "move away" from the scar; it is an early accreta, and the deeper it sits in the niche the more certainly it becomes increta or percreta. The verified outcome data make this concrete — in the Kaelin Agten/Timor-Tritsch series, pregnancies implanted in the niche almost all came to cesarean-hysterectomy for placenta increta or percreta, whereas those on the scar (with residual myometrium beneath them) mostly delivered normally, and a first-trimester myometrial thickness under 2 mm predicted a placenta accreta spectrum at delivery.

The other sites each have a characteristic failure mode:

• Cervical pregnancy implants in the endocervical canal below the internal os. The cervix is fibrous with almost no contractile muscle, so it cannot clamp down on the placental bed — disturbing it (a curette, a "missed miscarriage" evacuation) opens sinusoids that will not constrict, producing torrential, sometimes unstoppable, haemorrhage.
• Interstitial pregnancy implants in the intramural segment of the tube as it traverses the uterine cornu. That segment is surrounded by myometrium, so it is distensible and ruptures late — typically 7–12 weeks, by which time the conceptus is large and the cornual and ascending uterine vessels feed it, so rupture is a major obstetric haemorrhage, not a tubal trickle.
• Ovarian pregnancy implants in or on the ovary; the ovary is highly vascular and friable, and bleeding is usually early.
• Abdominal pregnancy implants on peritoneum, bowel, omentum or vessels; the placenta invades whatever it lands on, so the lethal moment is often at attempted removal of the placenta, not at diagnosis.
• Heterotopic pregnancy — a coexisting intrauterine and ectopic pregnancy — is the trap of the ART era: a positive scan showing an intrauterine sac falsely reassures, while the ectopic ruptures. Background incidence is roughly 1 in 30,000 spontaneously but of the order of 1 in 100–500 after IVF, so any assisted-conception pregnancy with pain, free fluid or an adnexal mass keeps heterotopic on the list even when the intrauterine pregnancy is seen.

One framework governs every management decision that follows: vascularity plus deficient myometrium equals danger, and the scar pregnancy is an early accreta.

Assessment

Getting the diagnosis: precise location, not "an ectopic"

The single most important investigation is a good transvaginal ultrasound that names the site, because "ectopic pregnancy" is not a diagnosis you can act on here — the treatment for a cervical pregnancy would kill in a scar pregnancy and vice versa. The mistakes are diagnostic before they are therapeutic.

The site-specific criteria a consultant must be able to state:

• Caesarean-scar pregnancy (SMFM criteria): an empty uterine cavity and empty endocervical canal; a gestational sac or placenta embedded in the hysterotomy scar; a triangular sac at ≤8 weeks (rounded/oval later) filling the niche; a thin (1–3 mm) or absent myometrial layer between sac and bladder; rich peritrophoblastic vascularity at the scar; and an embryo/yolk sac with or without cardiac activity. The discriminator from a cervical pregnancy and from a miscarriage in transit is that the sac sits anteriorly at the scar with the cavity and cervix empty.
• Cervical pregnancy: sac below a closed internal os, in the cervical canal, with cervical ballooning ("hourglass" uterus), trophoblastic flow on Doppler, and a negative "sliding sign" (the sac does not slide against the canal under probe pressure, unlike a miscarriage passing through).
• Interstitial pregnancy: an eccentric sac lying >1 cm lateral to the edge of the endometrial cavity, surrounded by a thin (<5 mm) myometrial mantle, with the "interstitial line sign" — an echogenic line from the cavity to the sac.
• Ovarian pregnancy: rarely diagnosed pre-operatively; classically a wide echogenic ring on the ovary. The historical Spiegelberg criteria define it pathologically — intact ipsilateral tube separate from the ovary, sac occupying the ovary, ovary connected to the uterus by the ovarian ligament, and ovarian tissue in the sac wall.
• Abdominal pregnancy: no myometrium surrounding the sac, sac separate from the uterus and tubes, with the placenta on a peritoneal surface; suspect it when the uterus is empty but a viable pregnancy is seen elsewhere.

MRI is the second-line problem-solver, not a screening tool. It earns its place when ultrasound cannot resolve a scar-versus-cervical question, when defining the depth of niche invasion and bladder involvement before surgery on a scar pregnancy, or when mapping placental and vascular relationships in an advanced abdominal pregnancy for theatre planning. It does not replace the transvaginal scan and should never delay treating an unstable woman.

Defining the site precisely — the classic exam distinctions

Three pairs of terms are routinely confused, and the confusion is dangerous:

• Interstitial versus cornual versus angular. Interstitial = implantation in the intramural (interstitial) segment of the tube — a true ectopic, outside the cavity, surrounded by myometrium. Cornual is properly reserved for a pregnancy in the rudimentary horn of a unicornuate (Müllerian-anomaly) uterus — a different entity with its own rupture risk, often needing horn excision. Angular is not an ectopic at all: it is an intrauterine pregnancy implanted in the lateral angle of the cavity, medial to the uterotubal junction and the round ligament — it can be viable and carried to term, though with a higher rate of malposition and abnormal placentation. Calling an angular pregnancy "interstitial" leads to destroying a viable intrauterine pregnancy; calling an interstitial pregnancy "angular" leads to a fatal cornual rupture. The discriminator is the position relative to the round ligament and the thickness and continuity of the surrounding myometrium.
• On the scar versus in the niche (caesarean-scar pregnancy): on the scar (endogenous/Type I) implants over the scar with residual myometrium beneath it and grows towards the cavity — better prognosis, occasionally compatible with a continuing pregnancy under intensive surveillance. In the niche (exogenous/Type II) implants deep in the dehiscent defect with little or no myometrium beneath, grows towards the bladder, and is the one that ruptures early and becomes percreta.

Staging the situation and the host

Once the site is named, the assessment is the same discipline as any acute gynaecological emergency, sharpened by the SA context. Quantify haemodynamic state (pulse, BP, shock index — a shock index >1 in a young woman is significant blood loss masked by physiological reserve). Take a quantitative serum hCG and, where the diagnosis is genuinely uncertain, enter the pregnancy-of-unknown-location (PUL) pathway: a single hCG cannot locate a pregnancy, and the old "discriminatory zone" (an intrauterine sac should be visible above an hCG of ~1500–2000 IU/L) is a caveat, not a rule — it varies with scanner, operator, and multiple pregnancy, and acting on a single threshold has caused viable intrauterine pregnancies to be given methotrexate. The validated modern approach is risk stratification (the M6 model logic): an initial progesterone and serial hCG triage women into probable failing PUL, probable intrauterine, or high-risk-for-ectopic, rather than betting on one number. Baseline FBC, group-and-save (cross-match if any instability), U&E and an HIV test complete the picture; in SA, confirm blood-product availability before you commit to an intervention that may bleed.

Management

Organise every answer immediate → ongoing → long-term, and let the over-arching principle drive it: stabilise, locate precisely, then choose a site-specific treatment from medical, surgical and expectant options — never a tubal-reflex treatment.

Immediate — make her safe

An unstable woman with a complicated ectopic is a surgical emergency regardless of site: resuscitate (two large-bore cannulae, crystalloid, activate the massive-transfusion protocol early because these bleeds are arterial and the scar/cervical/cornual sites bleed faster than a tubal rupture), cross-match, and get to theatre. The site dictates the operation, so the pre-operative scan is not a luxury — but instability overrides everything, and a laparotomy that controls the cornu or the abdominal placental bed comes before any elegant minimally invasive plan.

Ongoing — matching treatment to site (the stable patient)

The judgement is hardest in the stable woman, where medical, surgical and expectant routes genuinely compete and the right answer is site-specific. The honest backdrop is that this is cohort-and-consensus territory, not RCT territory — the only randomised evidence in non-tubal ectopic pregnancy is a handful of small trials, all in caesarean-scar pregnancy, so most "regimens" are extrapolated from case series and expert guidance.

A few threads need expanding, because the table compresses real decisions.

Methotrexate — which regimen, and why "single-dose" is a tubal habit that misleads in non-tubal disease. Systemic methotrexate is the inhibitor of dihydrofolate reductase that stops trophoblast proliferation; the SA/EML and international standard regimens are the single-dose (50 mg/m² IM, hCG checked day 4 and day 7, a second dose if hCG falls <15% between days 4–7), the two-dose, and the multi-dose (variable, MTX 1 mg/kg alternating with folinic acid 0.1 mg/kg) protocols. For a tubal ectopic the verified Cochrane evidence is that single-dose MTX is significantly less successful than laparoscopic salpingostomy (OR 0.38) while variable/multi-dose is not — and the verified two-dose-versus-single-dose meta-analysis shows the two-dose protocol succeeds more often (OR 1.84), with its advantage greatest exactly where single-dose is most tempting: high hCG (OR 3.23) and a large adnexal mass (OR 2.93). The same lesson carries into non-tubal disease: the higher the hCG and the larger the trophoblast, the weaker single-dose is, and the verified Lipscomb data show a prior ectopic and a high hCG independently predict single-dose failure. So systemic MTX, where it is used at all in non-tubal disease (interstitial, sometimes cervical or abdominal as an adjunct), should usually be a multi-/two-dose strategy — and for a caesarean-scar pregnancy systemic MTX alone is explicitly not recommended because the avascular scar tissue concentrates the drug poorly and the failure-then-haemorrhage sequence is dangerous; the role of methotrexate there is local, intragestational injection, often combined with a surgical or radiological technique.

Local injection — MTX or KCl. Under transvaginal ultrasound guidance, drug is injected directly into the gestational sac. KCl is the agent of choice when there is cardiac activity and you want to stop it fast (it is also the agent for the ectopic component of a heterotopic pregnancy, because it does not threaten the intrauterine pregnancy the way systemic MTX would); MTX locally targets the trophoblast. Local treatment puts a high drug concentration exactly where the avascular scar or fibrous cervix would otherwise see little systemic drug.

Uterine-artery embolisation as an adjunct, not a primary cure. UAE reduces the vascular supply to a scar or cervical pregnancy and is best used before or alongside evacuation/resection to control bleeding, or as rescue for haemorrhage. The verified Cochrane review of non-tubal ectopic found that, for caesarean-scar pregnancy, UAE/UACE before suction curettage gives lower blood loss than systemic MTX before curettage, though the certainty on success and complication rates is low. UAE is a tertiary-level resource with interventional radiology — a real access constraint in much of SA.

Managing massive haemorrhage is the same drill as any obstetric catastrophe but the bleeding sites are unforgiving: early activation of massive transfusion, tranexamic acid, mechanical tamponade (a Foley or balloon in the cervix/lower segment), uterine or internal-iliac artery ligation or embolisation, and — the decision a registrar avoids and a consultant must make calmly — hysterectomy as the definitive life-saving step when uterine conservation is losing the patient. Fertility is not worth a death.

Long-term — counselling, fertility and recurrence

Once she is safe, the long-term conversation is part of the management. Counsel on recurrence risk (a caesarean-scar pregnancy strongly predicts another and predicts accreta in any future pregnancy carried on that scar — many of these women should be counselled towards completed family/contraception, and SMFM explicitly recommends contraception counselling including long-acting and permanent methods), on the uterine-rupture risk in a future pregnancy after cornual or scar resection (plan an early scan to locate the next pregnancy, and counsel elective caesarean before labour), and on fertility — a salpingectomy for a heterotopic ectopic does not threaten the wanted intrauterine pregnancy, but a woman who has lost a tube, or had a cornual resection, needs realistic fertility counselling and early review in any subsequent pregnancy.

Guidelines compared

The bodies diverge, and where they fall silent is itself informative — the rarer the site, the more the "guideline" is really a consensus statement built on case series.

The recent, citable change is SMFM #63 (2022), which superseded Consult #49 and hardened the position against both expectant management and systemic-MTX-alone for caesarean-scar pregnancy — exactly the two things a generalist is most likely to do by tubal analogy.

The evidence & the controversy

The defining fact of this topic is how little randomised evidence exists. The verified Cochrane review of interventions for non-tubal ectopic pregnancy found only five randomised trials, all in caesarean-scar pregnancy, and no randomised trials at all for cervical, interstitial, ovarian or abdominal sites — those are managed entirely on cohort data and consensus. So the honest position is to manage from the SMFM consensus and the best cohort data, and to be explicit that the evidence base is observational. A "trial-proven" regimen for a cervical or interstitial pregnancy does not exist to be cited.

The live controversy is that the caesarean-scar pregnancy is a downstream harm of the rising caesarean-section rate. As caesarean rates climb worldwide and in South African public and private sectors, niche formation rises, and scar pregnancies and placenta accreta spectrum rise with them. This reframes a "rare ectopic" as an iatrogenic epidemic in slow motion, and it links the management of the index case to the prevention agenda: reducing the primary caesarean rate, and the wider conversation about caesarean on request, is the upstream answer to a problem that is currently treated only at its dangerous, expensive, downstream end. It is defensible to name this in an answer without moralising — present it as the epidemiological context, not a verdict on any individual delivery.

A second, sharper controversy is conservative continuation versus definitive treatment of a caesarean-scar pregnancy in a woman who wants the baby. The verified expectant-management data (Calì/D'Antonio) are sobering: a CSP with cardiac activity managed expectantly carried roughly a 10% first/second-trimester uterine-rupture risk, a 15% hysterectomy risk, and about a 75% rate of placenta accreta spectrum at delivery (mostly percreta) — yet a meaningful proportion reached the third trimester. SMFM recommends against expectant management precisely because of that morbidity, but also gives a delivery-timing recommendation (repeat caesarean at 34+0–35+6 weeks) for the woman who, fully informed, chooses to continue — a model of how a guideline can take a clear position and still respect autonomy. The defensible approach is a framework, not a decree: confirm the diagnosis and the niche depth, present the quantified risks honestly, involve a maternal-fetal-medicine and accreta-experienced team early, document the counselling, and support an informed decision either way.

Landmark trials & key evidence

A worked number: the early-diagnosis effect in caesarean-scar pregnancy. Severe composite morbidity was 5.9% when diagnosed at ≤9 weeks and 32.4% when diagnosed later — an absolute risk reduction of about 26.5%, so the number needed to diagnose-early to prevent one severe maternal-morbidity event is 1/0.265 ≈ 4 women. That is an enormous yield from nothing more than a first-trimester scan in a woman with a previous caesarean — which is precisely the argument for low-threshold early scanning of the prior-caesarean population in a high-caesarean-rate setting.

Worked viva — how to structure the answer

A typical stem: "A 34-year-old para 2, two previous caesareans, presents at 7 weeks with light bleeding. The transvaginal scan shows an empty uterine cavity, an empty cervical canal, and a gestational sac with a yolk sac and cardiac activity sitting anteriorly in the lower segment, with 2 mm of myometrium between the sac and the bladder." A high-scoring answer runs:

1. Name it precisely — "This is a caesarean-scar ectopic pregnancy implanted in the niche: empty cavity and cervix, sac in the hysterotomy scar, myometrium under 2 mm to the bladder. It is an early placenta accreta — the niche location and thin myometrium predict increta/percreta."
2. Frame the danger and stabilise — confirm she is haemodynamically stable, take hCG, FBC, group-and-save, and arrange this in a unit with theatre, blood and ideally interventional radiology, because the bleeding here is arterial.
3. State the plan against the guideline — "I would not manage this expectantly and I would not give systemic methotrexate alone — both are against SMFM #63. I would offer treatment now: operative resection (laparoscopic or transvaginal) or ultrasound-guided aspiration, with intragestational methotrexate or KCl to stop the cardiac activity, and uterine-artery embolisation available as a haemostatic adjunct. Sharp curettage alone is to be avoided."
4. Counsel on continuation honestly — "If she asked to continue the pregnancy, I would quote the expectant-management data: roughly a 10% rupture risk, 15% hysterectomy risk, and about 75% accreta spectrum at delivery, and I would involve a maternal-fetal-medicine/accreta team and document an informed, autonomous decision either way."
5. Justify from evidence — SMFM #63 for the management position, Kaelin Agten/Timor-Tritsch for the niche/accreta link and the <2 mm myometrium predictor, Calì for the expectant-management risks.
6. Close the loop — recurrence and accreta risk in any future pregnancy on this scar, contraception counselling including long-acting/permanent methods, and the public-health context of the rising caesarean rate.

Exam traps & red flags

• Treating a non-tubal ectopic with the tubal reflex. Single-dose systemic methotrexate or a quick salpingectomy is right for the tube and wrong for a scar, cervical, interstitial or abdominal pregnancy — each needs a site-specific plan.
• Systemic methotrexate alone for a caesarean-scar pregnancy. Explicitly recommended against (SMFM 1C) — the avascular scar concentrates the drug poorly and failure is followed by haemorrhage. The MTX role here is local/intragestational.
• Curetting a cervical or scar pregnancy blindly. Opens sinusoids that will not constrict — torrential, sometimes fatal, haemorrhage. Avoid sharp curettage alone.
• Confusing interstitial, cornual and angular. Interstitial = ectopic in the intramural tube; cornual = rudimentary-horn pregnancy; angular = an intrauterine, potentially viable pregnancy in the cavity angle. Mislabelling either destroys a viable pregnancy or misses a fatal cornual rupture.
• Forgetting heterotopic pregnancy after ART. A seen intrauterine sac does not exclude a coexisting ectopic when conception was assisted; pain/free fluid/adnexal mass keeps it on the list. Use KCl (not systemic MTX) for the ectopic if the intrauterine pregnancy is wanted.
• Trusting a single hCG / the discriminatory zone. It is a caveat, not a rule; act on serial hCG and validated PUL triage, not one threshold — a low single hCG does not exclude a scar or interstitial pregnancy.
• Underestimating interstitial rupture. It ruptures late (7–12 weeks) and bleeds heavily because cornual/uterine vessels feed it — a "small" ectopic by hCG can be a major haemorrhage.
• Peeling an adherent abdominal placenta. Removing a placenta off bowel or great vessels is the lethal step; leave it in situ when it cannot be removed safely and follow hCG.
• Missing the accreta future. A caesarean-scar pregnancy predicts accreta in the next pregnancy on that scar — failing to counsel on recurrence, future-pregnancy rupture and contraception is an incomplete answer.
• Overstating the evidence. Claiming a "trial-proven" regimen for cervical/interstitial/ovarian/abdominal disease, when no RCTs exist, reads as improvisation. Cite the consensus and the cohort data honestly.

Evidence anchors

• Cochrane — Interventions for non-tubal ectopic pregnancy (Long et al., 2020)
• Cochrane — Interventions for tubal ectopic pregnancy (Hajenius et al., 2007)
• SMFM Consult Series #63 — Cesarean scar ectopic pregnancy (Am J Obstet Gynecol 2022)
• Kaelin Agten et al. — CSP "on the scar" vs "in the niche" (Am J Obstet Gynecol 2017)
• Timor-Tritsch et al. — CSP outcome by gestational age at diagnosis (Eur J Obstet Gynecol Reprod Biol 2021)
• Calì et al. — Outcome of expectantly managed CSP (Ultrasound Obstet Gynecol 2018)
• Alur-Gupta et al. — Two-dose vs single-dose methotrexate meta-analysis (Am J Obstet Gynecol 2019)
• Lipscomb et al. — Previous ectopic as predictor of methotrexate failure (Fertil Steril 2004)
• Bobdiwala et al. — M6 model two-step PUL triage (Ultrasound Obstet Gynecol 2020)
• Poole et al. — Early abdominal ectopic pregnancy systematic review (Gynecol Obstet Invest 2012)
• RCOG Green-top Guideline No. 21 — Diagnosis and Management of Ectopic Pregnancy.
• NICE NG126 — Ectopic Pregnancy and Miscarriage: diagnosis and initial management (2019).
• ACOG Practice Bulletin 193 — Tubal Ectopic Pregnancy (2018) — single-, two- and multi-dose methotrexate protocols and eligibility.
• South Africa NDoH — National Integrated Maternal and Perinatal Care Guidelines, 5th edition (2024) — referral pathways and EML drug availability.