Manage termination of pregnancy and early pregnancy loss, including methods, complications and the South African medico-legal framework

In one line

Termination of pregnancy and early pregnancy loss are managed with the same small toolkit — mifepristone, misoprostol and uterine evacuation — and in South Africa the consultant's job is as much to make a legal, equitable, complication-light service work under the Choice on Termination of Pregnancy Act as it is to pick the right regimen; the single most important principle is that safe abortion is harm reduction, and obstructing access (through delay, conscientious objection without referral, or under-trained providers) is itself a cause of maternal death.

This chapter assumes the early-pregnancy groundwork — dating, the discriminatory zone, viability criteria — and addresses running and defending a service. The mechanics of medical evacuation overlap almost entirely between an induced abortion and a spontaneous miscarriage, so the two are covered together; where the management forks it is on consent and legality, not on pharmacology. Rhesus prophylaxis is covered in depth at rhesus-alloimmunisation and the haemorrhage pathway at postpartum-haemorrhage; recurrent loss has its own dedicated work-up and is not re-derived here.

Mechanism & pathophysiology

Two drugs and one piece of physiology underpin the whole toolkit: a pregnancy is maintained by progesterone acting on the decidua and myometrium, and the cervix is a collagen gatekeeper that must soften before the uterus can empty without trauma.

Mifepristone is a competitive antiprogestin. It binds the progesterone receptor with higher affinity than progesterone itself but produces no agonist signal, so the decidua loses its progesterone support, undergoes necrosis and detaches; the trophoblast separates, and prostaglandin-degrading enzyme activity falls. Two further effects matter clinically: mifepristone softens and ripens the cervix, and it up-regulates myometrial prostaglandin receptors and gap junctions, priming the uterus so that a subsequent prostaglandin produces stronger, better-coordinated contractions. This is why mifepristone is given first and then a deliberate interval is left before the prostaglandin — the pre-treatment is what converts a sluggish, painful misoprostol-only evacuation into an efficient one.

Misoprostol is a synthetic prostaglandin E1 analogue. It binds myometrial EP2/EP3 receptors to drive contractions and, independently, breaks down cervical collagen to ripen the cervix. It is heat-stable, cheap, orally/vaginally/sublingually/buccally active and needs no cold chain — the single most important drug in a low-resource abortion and post-abortion service. Route changes the pharmacokinetics: oral gives a fast, short peak (more GI side-effects), while vaginal, sublingual and buccal give a slower, sustained exposure with higher bioavailability and are preferred for evacuation. Given alone it works, but less reliably and more slowly than when the uterus has first been primed by mifepristone — the entire logic of the combination regimen.

The same physiology drives a spontaneous miscarriage: progesterone withdrawal (here from a failing trophoblast) triggers decidual breakdown and prostaglandin-mediated contractions. So a missed or incomplete miscarriage is, mechanistically, a pregnancy that has already begun the antiprogestin cascade on its own — which is exactly why the same drugs complete it, and why mifepristone pre-treatment helps a missed miscarriage just as it helps an induced abortion.

The South African legal framework

In an SA exit exam this topic is examined as much on the Choice on Termination of Pregnancy Act 92 of 1996 (as amended by Act 1 of 2008) as on clinical method. The Act is one of the most liberal abortion laws in the world, and a consultant must be able to state its thresholds precisely.

Four provisions reliably generate the curveballs:

• Consent is the woman's alone. No spousal, partner or parental consent is required at any gestation. A husband cannot veto, and cannot demand, a termination.
• Minors. A girl of any age who has the capacity for informed consent may consent to her own termination without parental assistance — this is the CTOP Act read together with the Children's Act 38 of 2005 (which sets the general medical-consent floor at 12 years with capacity). She must be advised to consult her parents, guardian or family, but a termination may not be refused or delayed if she chooses not to involve them. "Bring your mother first" is unlawful gatekeeping.
• The grounds in 13–20 weeks are stated, not adjudicated. The woman states the ground (e.g. socio-economic circumstances); the provider does not sit as a tribunal investigating its truth. The rape ground in particular does not require a police report, a J88, or proof — demanding any of these as a precondition is both unlawful and a recognised access barrier.
• Conscientious objection. A provider may decline to perform a termination on grounds of conscience, but the right is narrow: it does not extend to refusing emergency care, does not cover pre- or post-abortion counselling and care, and — critically — carries a duty to refer the woman without delay to someone who will help. An objector who obstructs, stalls or fails to refer is not exercising a protected right; they are creating the delay that pushes a first-trimester request into the second trimester (or out of the legal service altogether), and that delay is the mechanism of harm. Conscientious objection and the woman's right of access are reconciled by the referral obligation — the objector steps aside but does not block the door.

The gap between this permissive law and the service reality is the central SA story: a minority of designated facilities actually provide terminations, second-trimester services are concentrated and often distant, and the consequence is that women present later, travel further, or use unregulated misoprostol bought informally. The unsafe-abortion deaths the CTOP Act was written to prevent persist not because the law is restrictive but because access is patchy — which reframes "manage TOP" as a service-delivery and equity problem, not only a prescribing one.

Assessment

The pre-procedure assessment is deliberately lean — over-investigating is itself a barrier — but a consultant must know what genuinely changes management.

• Confirm and date the pregnancy. Gestational age sets both the legal pathway and the regimen. Ultrasound dating is ideal but not mandatory for early medical abortion where dates are certain and there are no red flags — WHO is explicit that routine scanning is not a prerequisite, and insisting on it where it is unavailable simply delays care. A scan is warranted when gestation is uncertain, ectopic is plausible (pain, prior ectopic, IUCD, suspicious bleeding pattern), or molar disease is suspected.
• Exclude ectopic before attributing bleeding/pain to an intrauterine process. A misoprostol regimen will not treat — and may mask — an ectopic. Where the pregnancy is of unknown location, follow the ectopic pathway rather than presuming a miscarriage or proceeding to evacuation; advanced and non-tubal ectopics carry their own management considerations beyond the scope of this objective.
• Rhesus status — and the modern reframing. Determine blood group, but appreciate that the indication for anti-D in early pregnancy has narrowed sharply (developed in the trials section). It is no longer a reason to delay an early medical abortion.
• Bloods are minimal. Haemoglobin where anaemia or significant bleeding is likely; no routine baseline panel is required for uncomplicated early abortion. HIV testing is offered as part of SA care, but a result is not a gate to the procedure.
• Screen the context, not just the uterus. Contraceptive intentions (so a LARC can be placed at the same visit), intimate-partner violence and coercion (is this her choice?), and — for a minor or a disclosed rape — the safeguarding and forensic pathways. The clinical encounter is often the only contact point, so the assessment is also a screening opportunity.

For early pregnancy loss, the assessment adds a viability/completeness judgement: a missed miscarriage (non-viable intrauterine pregnancy, closed os), an incomplete miscarriage (retained tissue, open os, ongoing bleeding), or a complete miscarriage. Resist over-reliance on a single number here — endometrial thickness does not reliably distinguish retained products from normal decidua (Sawyer 2007 found no useful cut-off), so a "thick endometrium" on its own is not an indication to operate. Manage the woman — her bleeding, pain and preference — not a millimetre on the report.

Management

Organise it immediate → ongoing → long-term, and let method follow gestation and the woman's choice rather than provider habit.

Immediate — choosing and delivering the method

Medical abortion, under 12 weeks (the workhorse). The combination regimen is mifepristone 200 mg orally, then 24–48 hours later misoprostol 800 µg by the vaginal, sublingual or buccal route. This is the regimen behind ACOG Practice Bulletin 225 (medication abortion up to 70 days) and the WHO guideline. Where mifepristone is unavailable — a real SA constraint — misoprostol alone (typically 800 µg vaginally/sublingually, repeated 3-hourly for up to three doses) is effective but slower, more painful and with a higher continuing-pregnancy rate; mifepristone pre-treatment is the upgrade wherever it can be sourced. Complete abortion rates with the combination exceed 95%.

Medical abortion, 12 weeks and beyond. The regimen shifts to mifepristone 200 mg, then (≥24 hours later) misoprostol 400 µg buccally/sublingually/vaginally, repeated every 3 hours until expulsion (WHO Recommendation 29). The interval and the repeated lower dose reflect a uterus that is now larger and more prostaglandin-sensitive. Beyond ~24 weeks the misoprostol dose is reduced and clinical judgement governs, because the data thin out and uterine sensitivity rises.

Surgical abortion. Manual vacuum aspiration (MVA) is the first-trimester (and up to ~14-week) surgical method — a hand-held syringe, no general anaesthetic, fast, and ideally suited to district and primary level; it has replaced sharp curettage, which is more morbid (Asherman's, perforation) and should not be the default. From ~14 weeks the surgical method is dilatation and evacuation (D&E), which requires cervical preparation first (misoprostol and/or mifepristone, or osmotic dilators) and specific operator skill, so it is a referral-level procedure. Cervical priming before any surgical abortion reduces cervical injury and is standard from the early second trimester (and considered earlier on judgement).

The choice between medical and surgical at a given gestation is largely the woman's, once she understands the trade-offs: medical avoids instrumentation and can be partly self-managed but means cramping, bleeding and a less predictable timeline; surgical is quick and definitive but instrumented. Neither is "better" in the abstract — autonomy decides.

Immediate — early pregnancy loss runs on the same toolkit

The same three options apply to miscarriage, and the evidence (below) lets you counsel each honestly:

• Expectant management — appropriate for incomplete miscarriage and acceptable for some missed miscarriages; safe, but slower and with a meaningful crossover to unplanned surgical evacuation. Reasonable where bleeding is settling and the woman accepts uncertainty.
• Medical management — misoprostol, and for a missed miscarriage, mifepristone pre-treatment then misoprostol (the MifeMiso upgrade). The misoprostol dose for early pregnancy loss is gestation- and scenario-specific (lower than for a viable termination because the pregnancy has already failed).
• Surgical management — MVA under local anaesthetic, the SA-appropriate default for incomplete miscarriage with heavy bleeding or where the woman wants it over quickly; faster and more complete than expectant/medical, at the cost of instrumentation.

The unifying message MIST gave us is that all three are safe — infection rates do not differ by method — so the decision is driven by completeness, speed and the woman's preference, not by a fear that one route is dangerous.

Ongoing — the immediate after-care

• Analgesia — NSAIDs are first-line and do not reduce abortion efficacy; offer them routinely rather than reactively.
• Confirm completion — for medical abortion this is by symptoms plus, where needed, a low-sensitivity urine pregnancy test at ~2–4 weeks or a falling hCG; routine same-week scanning is unnecessary and over-diagnoses "retained products" on the basis of an unreliable endometrial measurement. Persistent heavy bleeding, fever or a continuing pregnancy are the genuine triggers to act.
• Anti-D — give per the current threshold (see trials section): in practice, not for early medical management, and individualised for surgical evacuation per gestation.

Complications — recognise, grade, and treat the predictable ones

Abortion and miscarriage management share a small, predictable complication set, and a consultant is expected to grade and pre-empt them rather than merely list them.

• Haemorrhage. The commonest serious complication. Most bleeding after medical management is expected heavy menstrual-type loss; the pathological signals are soaking more than two pads an hour for two hours, clots, or haemodynamic change. Causes are atony, retained products, cervical/uterine injury, and — rarely — a missed ectopic or undiagnosed molar/coagulopathy. Manage on the haemorrhage ladder (uterotonics — misoprostol/oxytocin, evacuation of retained tissue, tamponade, escalate), exactly as at postpartum-haemorrhage; the principle of acting early on a rising shock index applies just as it does intrapartum.
• Retained products of conception. Incomplete evacuation causes ongoing bleeding, pain and infection. Diagnose on the clinical picture (persistent bleeding, open os, tissue) supported by ultrasound — but, as above, not on endometrial thickness alone — and treat with repeat misoprostol or MVA. Over-diagnosis (operating on a "thick stripe" in a settling, well woman) is as much an error as under-diagnosis.
• Infection and post-abortal sepsis. Endometritis presents with fever, pain, offensive discharge and a tender uterus; untreated it progresses to pelvic sepsis. Treat promptly with broad-spectrum antibiotics and evacuate retained tissue (source control). Post-abortal sepsis with a closed, retained, infected uterus is a surgical emergency — antibiotics alone will not control an infected cavity. MIST reassures that baseline infection risk does not differ by method; it does not license ignoring sepsis when it appears.
• Uterine perforation. A surgical-evacuation risk, higher with sharp curettage, an acutely retroverted or scarred uterus, and inexperienced operators — which is the practical argument for MVA, cervical priming and adequate training. A suspected perforation with peritonism, broad-ligament haematoma or visceral injury needs laparoscopy/laparotomy; a quiet fundal perforation in a stable woman may be managed conservatively with observation.
• Cervical injury. Laceration from forced dilatation; prevented by cervical priming (misoprostol/mifepristone/osmotic dilators) before instrumentation — the single most effective preventive step, and the reason priming is built into D&E.
• Failed or ongoing pregnancy. Medical regimens have a small failure rate (higher for misoprostol-only); a continuing pregnancy after a completed abortion attempt must be recognised and offered surgical completion, with explicit counselling about misoprostol's teratogenic (Möbius-sequence/limb) signal if the pregnancy were to be continued — a point that occasionally surfaces in vivas.
• Rhesus sensitisation. Real but, in early pregnancy, far smaller than historically assumed — the basis for the narrowed anti-D indication discussed below; covered fully at rhesus-alloimmunisation.

Long-term — contraception and the harm-reduction frame

The post-abortion contraception conversation is central, not a discharge afterthought, because the period immediately after an abortion is one of high fertility and high motivation. Immediate LARC — a copper or levonorgestrel IUD placed at the time of surgical evacuation, or an implant inserted on the same day as medical abortion — is the single most effective intervention against repeat unintended pregnancy, and the evidence is that immediate placement beats "come back in six weeks" (most women do not). Frame the whole episode as harm reduction: a woman who has accessed a safe, legal termination and left with effective contraception has been served well; one turned away to seek unsafe means has not.

Guidelines compared

The international consensus on method is strong; the live differences are on who provides, what testing is mandatory, and anti-D.

The direction of travel is uniform — decentralise, de-medicalise the unnecessary tests, and stop giving anti-D where it adds nothing — and SA's law is already permissive enough to follow it; the lag is in implementation, not in policy.

The evidence & the controversy

Three threads define the modern, exam-worthy debate.

First, the anti-D reversal. For decades RhD-negative women were given anti-D after any abortion or miscarriage. The recent re-appraisal (WHO 2022, NICE, and the ACOG 2024 clinical practice update with a concordant SMFM statement) concluded that the volume of feto-maternal haemorrhage before ~10–12 weeks is too small to credibly cause sensitisation, the historical evidence base was thin, and routine anti-D in early pregnancy is not warranted — for medical management certainly, and increasingly for surgical too. This matters in SA both as a cost/access argument (anti-D is a scarce, expensive blood product) and as a barrier-removal argument (waiting for blood-grouping should never delay an early medical abortion). The defensible position states the new threshold, gives the FMH-volume rationale, and acknowledges that practice is still catching up — rather than reverting to the old "all RhD-negative women get anti-D."

Second, decentralisation, self-management and telemedicine. WHO's 2022 guideline made the boldest move: early medical abortion can be safely self-managed by the pregnant person, and provision task-shared to non-physician cadres. The COVID-era expansion of telemedicine "no-test, no-touch" medical abortion (mifepristone and misoprostol posted to the woman after a remote consultation) was studied at scale and found safe and effective, and has reshaped abortion delivery in several countries. For SA the opportunity is obvious — it could route around the access gap created by too few facilities and conscientious objectors — but the controversy is real: safeguarding (coercion, gestational-age certainty without a scan, the rare ectopic missed remotely) has to be designed in, not waved away. The balanced position is that self-management plus telemedicine is a genuine equity lever with a triage-and-safety net, not a free-for-all.

Third, the harm-reduction versus access politics. Abortion is one of the few areas where the law is settled and liberal in SA but the service is contested on the ground — through facility refusal, provider objection, and stigma. The current, defensible professional framing is harm reduction: unsafe abortion remains a contributor to maternal mortality, every barrier (delay, unnecessary test, un-referred objection) pushes women toward it, and the clinician's ethical job is to lower barriers within the law. Where this becomes genuinely difficult — a conscientiously objecting registrar, a minor wanting confidentiality from her parents, a second-trimester request at a facility without a willing provider — the resolution is a framework, not a verdict: name the competing principles (the woman's autonomy and right of access; the provider's conscience; the duty of non-maleficence that includes not obstructing), and resolve them through the referral obligation and the legal thresholds rather than by imposing the clinician's own values.

A short topical thread worth holding: misoprostol bought informally — online or over a counter — is now a significant route by which SA women self-abort outside the formal service. Rather than treat every such woman as a culprit, the harm-reduction stance is to provide non-judgemental assessment, complete an incomplete process safely, give contraception, and recognise that the informal market is a symptom of the access gap the formal service should be closing.

Landmark trials & key evidence

A worked arithmetic from MifeMiso: the absolute reduction in surgical intervention is 25% − 17% = 8 percentage points, so the number needed to treat with mifepristone pre-treatment to avoid one surgical evacuation ≈ 1/0.08 ≈ 13 — a strong yield from one cheap oral tablet, which is the core of the argument for adding mifepristone to misoprostol for missed miscarriage wherever it can be obtained.

Worked viva — the conscientious-objection / rape-disclosure curveball

A typical SA stem: "A 15-year-old presents alone requesting a termination at 11 weeks. She discloses the pregnancy resulted from rape but does not want to involve the police or her parents. The on-call registrar says he conscientiously objects to abortion." A high-scoring answer runs:

1. State the legality clearly. "At 11 weeks she is within the on-request period, so no ground or proof is needed at all — the rape disclosure does not even have to be adjudicated here. A minor with capacity may consent herself; she must be advised to involve a trusted adult but cannot be compelled to, and care may not be withheld if she declines. No police report or J88 is a precondition for the termination."
2. Handle the objection within its limits. "The registrar may decline to perform the procedure, but his right does not let him obstruct or refuse to help — he has a duty to refer her without delay to a willing provider, and the objection does not cover counselling, emergency care or post-abortion care. I would arrange a willing provider immediately so there is no delay; a delay that pushed her toward 13 weeks, or out of the service, would be the actual harm."
3. Run the parallel safeguarding pathway without letting it gate the abortion. "Separately and in parallel — not as a precondition — I would offer the forensic and protective response a rape disclosure in a minor triggers: sensitive history, offer of forensic care and a Thuthuzela/GBV referral, HIV and STI post-exposure assessment, emergency-contraception logic moot here, and the child-protection obligations. These run alongside, never ahead of, her care."
4. Complete the care. "Early medical abortion (mifepristone 200 mg then misoprostol 800 µg), NSAID analgesia, anti-D not indicated at this gestation, and — central, not optional — immediate effective contraception (an implant or IUD) before she leaves, with follow-up and psychological support."
5. Name the framework, not a verdict. "The competing principles are her autonomy and right of access, the registrar's conscience, and non-maleficence — which here includes not obstructing. They reconcile through the referral duty and the legal thresholds, which let the objector step aside without closing the door."

Exam traps & red flags

• Demanding a reason, a police report or proof for the rape ground. Under 12 weeks it is on request; even in 13–20 weeks the ground is stated, not investigated. Requiring a J88 or police docket is unlawful gatekeeping and a recognised cause of delay.
• Requiring parental consent for a minor. A girl with capacity consents herself at any age; she is advised to involve family but may not be refused if she declines. "Bring your mother" is the wrong answer.
• Spousal or partner consent. Never required — not to permit and not to refuse.
• Conscientious objection without referral. Declining to perform is permitted; obstructing, stalling, refusing post-abortion/emergency care, or failing to refer is not — that is the trap that converts an objection into a harm.
• Insisting on a scan or full bloods before early medical abortion. Over-investigation is a barrier; routine ultrasound and lab panels are not prerequisites where dates are certain and there are no red flags.
• Treating bleeding/pain as miscarriage without excluding ectopic. Misoprostol will not treat an ectopic and may delay its diagnosis — exclude it first where the location is unknown.
• Using endometrial thickness as the trigger to evacuate. No cut-off reliably distinguishes retained products from decidua (Sawyer 2007); manage symptoms and bleeding, not a millimetre.
• Reflex anti-D for every RhD-negative early abortion/miscarriage. The threshold has moved — not routinely below ~10–12 weeks for medical management (WHO/NICE/ACOG 2024); waiting on a blood group should not delay an early medical abortion.
• Sharp curettage as the default surgical method. MVA is the standard for first-trimester evacuation — sharp curettage carries more perforation and Asherman's risk.
• Forgetting contraception. Discharging without immediate LARC (or at least a concrete plan) wastes the highest-motivation, highest-fertility moment and predicts a repeat unintended pregnancy.
• Misoprostol-only when mifepristone was available. Combination is more effective and less painful; defaulting to misoprostol-only where mifepristone could have been sourced is under-treatment.

Evidence anchors

• MIST trial — Trinder et al., management of miscarriage (expectant/medical/surgical), BMJ 2006
• MifeMiso trial — Chu et al., mifepristone + misoprostol vs misoprostol alone for missed miscarriage, Lancet 2020
• WHO Abortion care guideline (2022; second edition 2024)
• Endometrial thickness/volume for incomplete miscarriage — Sawyer et al., Ultrasound Obstet Gynecol 2007
• South African Choice on Termination of Pregnancy Act 92 of 1996 (as amended by Act 1 of 2008) — on request to 12 weeks (doctor or trained midwife/nurse), 13–20 weeks on stated grounds (doctor), restricted after 20 weeks; woman's consent only; minors with capacity consent themselves; conscientious objection with a duty to refer.
• South African Children's Act 38 of 2005 — medical-treatment consent from age 12 with capacity (read with the CTOP Act for minors seeking termination).
• NICE NG126 — abortion / ectopic pregnancy and miscarriage: anti-D not routine for medical management ≤10 weeks; mifepristone pre-treatment for missed miscarriage.
• ACOG Clinical Practice Update (2024) — Rh D immune globulin after abortion or pregnancy loss at less than 12 weeks; concordant SMFM statement.
• ACOG Practice Bulletin 225 — Medication Abortion Up to 70 Days of Gestation (mifepristone 200 mg + misoprostol 800 µg).