Manage postpartum venous thromboembolism and secondary postpartum haemorrhage

In one line

The puerperium is the single highest-risk period for venous thromboembolism in a woman's reproductive life, so every postnatal woman earns a documented VTE risk assessment and most need prophylactic low-molecular-weight heparin for at least 10 days; secondary postpartum haemorrhage — excessive bleeding from 24 hours to 12 weeks after delivery — is endometritis and retained products until proven otherwise, but the trap is the woman whose bleeding is actually a uterine artery pseudoaneurysm, in whom a curette makes a survivable problem catastrophic.

These two complications are linked by their timing and by a shared clinical discipline: both kill quietly, days after the woman has gone home, and both are managed badly when the early signal is read as trivial.

Mechanism & pathophysiology

Venous thromboembolism in the puerperium is Virchow's triad maximally expressed. Pregnancy is a physiologically prothrombotic state engineered to survive the haemostatic challenge of placental separation, and that protection does not switch off at delivery — it peaks. Hypercoagulability: oestrogen-driven rises in fibrinogen and factors VII, VIII and X, a fall in protein S, and acquired resistance to activated protein C tip the balance towards clot, and these changes persist for roughly six weeks postpartum before returning to baseline. Stasis: venous capacitance and the mechanical compression of the gravid uterus on the iliac veins — worse on the left, because the left iliac vein is crossed by the right iliac artery — slow flow, which is why iliofemoral DVT in pregnancy is overwhelmingly left-sided and often presents with whole-leg rather than calf swelling. Endothelial injury: delivery itself, and caesarean section in particular, traumatises pelvic vasculature and activates the endothelium. The puerperium is when all three are simultaneously maximal, which is why the daily risk of VTE is higher in the weeks after birth than at any point in the antenatal period, and the first week is the most dangerous of all.

This matters because VTE remains a leading direct cause of maternal death across every high-quality confidential enquiry, and almost all of those deaths are from pulmonary embolism in women who were either never risk-assessed or were under-dosed and under-treated. The disease is largely preventable; the deaths are largely failures of a system that did not weigh the risk and act.

Two consequences of this physiology shape every downstream decision. First, the prothrombotic milieu does not respect the artificial line between antenatal and postnatal care, so a woman with antenatal risk factors is more, not less, at risk once delivered — the assessment must be repeated after birth, not assumed to have ended with it. Second, because the changes take roughly six weeks to resolve, prophylaxis pitched to the inpatient stay alone — a few days after a vaginal birth, a little longer after caesarean — leaves the larger part of the high-risk window uncovered. The pathophysiology is the argument for community-prescribed, weeks-long prophylaxis in the women who need it.

Secondary postpartum haemorrhage is a disorder of the involuting placental bed. After delivery the placental site is a raw, vascular wound that heals by thrombosis of the spiral arteries and progressive involution of the uterus. Three things derail that process. Retained products of conception keep the bed open: residual trophoblast and decidua prevent the uterus from contracting down on the placental site and act as a nidus for infection. Endometritis — ascending infection of the decidualised endometrium — inflames the bed, impairs involution and is itself frequently driven by retained products, so the two travel together. Subinvolution of the placental site is failure of the normal thrombosis-and-obliteration of the utero-placental vessels, which then re-bleed days to weeks later even when the cavity looks empty.

The dangerous minority is vascular. A uterine artery pseudoaneurysm is a contained rupture of an arterial wall — classically after the uterine trauma of caesarean section — in which blood is held only by a thin fibrous capsule and surrounding clot rather than a true three-layered vessel wall. Because the wall is incomplete, the lesion enlarges with arterial pressure and ruptures unpredictably, which is why the bleeding it causes is characteristically sudden, heavy and intermittent — quiescent while the capsule holds, then catastrophic when it gives. An arteriovenous malformation is an abnormal direct communication between uterine arteries and veins, either congenital or, more often in this setting, acquired after caesarean, evacuation or other instrumentation; it presents similarly with recurrent heavy bleeding and a uterus that looks structurally normal on greyscale imaging. Both are destroyed by the instrument that treats retained products: passing a curette into a pseudoaneurysm tears the capsule and converts intermittent bleeding into exsanguination. Recognising that a vascular cause can masquerade as ordinary retained products is the whole game.

The link between the two vascular causes and the rest of the topic is the caesarean section. The same operative trauma that injures pelvic veins and drives postpartum VTE also injures the uterine arteries and seeds pseudoaneurysm and AVM — so as caesarean rates rise, both these complications become commoner, and the woman who had a caesarean is simultaneously the woman who most needs thromboprophylaxis and the woman in whom secondary bleeding must never be reflexively curetted. Finally, gestational trophoblastic disease — persistent trophoblast after a molar or even a non-molar pregnancy — can present as secondary PPH and is the reason a βhCG belongs in the workup of bleeding that does not behave (see gestational-trophoblastic-disease).

Assessment

Postpartum VTE — assess every woman, suspect clinically, image decisively.

• The risk assessment is the intervention. Every woman has a documented VTE risk assessment at booking, repeated after any admission, and — critically — again after delivery before discharge. Postnatal risk factors are cumulative: caesarean section (especially emergency), class-3 obesity (BMI ≥40), age >35, parity ≥3, smoking, pre-eclampsia, prolonged labour, mid-cavity instrumental delivery, postpartum haemorrhage, immobility, infection, and any personal or strong family history of VTE or known thrombophilia. The number of risk factors drives the decision, so the assessment is not paperwork — it is the treatment decision.
• Suspected DVT. Unilateral (usually left) leg pain and swelling, sometimes whole-leg, sometimes only lower abdominal or buttock pain if the thrombus is iliac. Compression duplex ultrasound is the primary test. If it is negative but clinical suspicion is high, stop anticoagulation and repeat the scan on days 3 and 7 — a negative early scan does not exclude a propagating iliac clot.
• Suspected PE. Breathlessness, pleuritic chest pain, tachycardia, haemoptysis, collapse. An ECG and chest X-ray come first (fetal/maternal radiation from a CXR is negligible and it identifies pneumonia, pneumothorax or other causes). If there are leg symptoms, a duplex showing DVT confirms the need to treat and spares chest imaging. Without leg symptoms, choose V/Q scan or CTPA; if the CXR is abnormal, CTPA is preferred. The choice is a genuine consent conversation: compared with CTPA, V/Q carries a marginally higher childhood-cancer risk to the child of a future pregnancy, while CTPA delivers a higher radiation dose to maternal breast tissue with its attendant lifetime breast-cancer risk. There is no free option; the woman should be part of the decision.
• The investigation that is wrong in pregnancy. D-dimer is not used to investigate acute VTE in pregnancy or the puerperium. It is physiologically elevated by pregnancy and rises further with delivery, surgery and any inflammation, so a "positive" result is meaningless and a "negative" result cannot be relied upon to rule out. Treating a pregnant woman's leg or chest symptoms on a D-dimer is a category error that either floods imaging with false positives or, worse, falsely reassures.

Secondary PPH — date it, infection-screen it, image it sceptically, and never forget the vascular cause.

• History and examination. When did the bleeding start, how heavy, is it intermittent or continuous, was there a sudden gush? Fever, offensive lochia, abdominal pain and uterine tenderness point to endometritis; an open cervical os and tissue point to retained products. Quantify the haemodynamic state — pulse, blood pressure, conscious level — because the woman bleeding intermittently from a pseudoaneurysm can be deceptively well between gushes and then crash. Take note of the delivery: a caesarean shifts the probabilities towards endometritis and pseudoaneurysm and away from retained products.
• Infection screen. High vaginal and endocervical swabs, FBC, and cultures if systemically unwell. A low yield of positive swabs does not exclude endometritis; treat on the clinical picture.
• Ultrasound, read with its limits in mind. Pelvic ultrasound may help exclude retained products, but the ultrasound diagnosis of retained products is unreliable — reported sensitivity ranges 44–94% and specificity 16–92%. An echogenic mass with a thickened cavity favours retained products, but a clean-looking cavity does not exclude them and a "mass" may be clot. Colour-flow Doppler earns its place here: a focal area of turbulent, high-velocity flow is the signature of a pseudoaneurysm or AVM, and seeing it changes the plan from theatre to interventional radiology.
• βhCG when the bleeding does not behave. Persistent or recurrent secondary bleeding, particularly after a molar or non-viable pregnancy, demands a βhCG to exclude gestational trophoblastic disease.
• Read the pattern, not just the snapshot. The two commonest causes usually announce themselves — endometritis with fever, pain, tender bulky uterus and offensive lochia; retained products with an open os, tissue and steady bleeding — and they frequently coexist. What separates the consultant from the trainee is holding the third possibility open when the picture does not fit: bleeding that is dramatic and intermittent, in a woman who is afebrile with a near-empty cavity, especially after a caesarean, is a vascular cause until Doppler says otherwise. The cervical os, the temperature chart, the delivery mode and the tempo of the bleeding together point at the cause more reliably than any single ultrasound image.

Management

Postpartum VTE. Structure it as immediate → ongoing → long-term.

Immediate — prophylaxis. The postnatal risk assessment dictates LMWH. The thresholds, in weight-appropriate prophylactic doses:

Prophylactic dosing is by booking or most-recent weight: enoxaparin 20 mg daily if <50 kg, 40 mg daily for 50–90 kg, 60 mg daily for 91–130 kg, 80 mg daily for 131–170 kg, and 0.6 mg/kg/day above 170 kg (dalteparin and tinzaparin have equivalent weight bands). The single most common error is to dose by a default 40 mg in a woman who weighs 110 kg — under-dosing the highest-risk patient. The timing of the first dose around delivery is a safety detail worth stating: prophylactic LMWH is held around the time of regional anaesthesia and delivery and started once haemostasis is secure and a safe interval has elapsed after a spinal or epidural catheter is sited or removed, and is deferred while the woman is actively bleeding — because the prize of preventing VTE must not be bought with a spinal haematoma or a worsened postpartum haemorrhage. Mechanical measures (early mobilisation, hydration, and graduated compression stockings where there is a contraindication to or in addition to LMWH) complement, but do not replace, pharmacological prophylaxis in the high-risk woman.

Immediate — treatment of confirmed VTE. Therapeutic LMWH is the anticoagulant of choice, titrated to booking/early-pregnancy weight (enoxaparin 1 mg/kg twice daily, or 1.5 mg/kg once daily). Start treatment as soon as VTE is suspected and continue until objective testing excludes it, unless anticoagulation is strongly contraindicated — you do not wait for the scan to anticoagulate a breathless, tachycardic puerperal woman.

Ongoing — duration and the postnatal switch. Therapeutic anticoagulation continues for at least six weeks postnatally and for a minimum of three months in total. Postnatally the woman is offered a choice: continue LMWH, or switch to an oral anticoagulant. Warfarin may be started postpartum (avoided until at least day 5 because of the bleeding and the transient procoagulant dip), with the inconvenience of INR monitoring. The breastfeeding rule is the one to be certain of: neither LMWH nor warfarin is contraindicated in breastfeeding — both are safe. The direct oral anticoagulants (rivaroxaban, apixaban) are avoided in lactation: there are essentially no safety data on their excretion into breast milk, so they are not a breastfeeding-compatible option, and they are also avoided antenatally. Warfarin is teratogenic and is not used antenatally, but postpartum it is a legitimate breastfeeding-safe choice.

Massive, life-threatening PE. The collapsed, shocked puerperal woman with suspected massive PE is a resuscitation, managed by a multidisciplinary team. Intravenous unfractionated heparin is the preferred initial anticoagulant in massive PE with haemodynamic compromise — it is titratable and rapidly reversible, which matters in a woman who may bleed. Arrange CTPA within an hour; if massive PE is confirmed (or, in extremis, before confirmation), thrombolysis is considered despite recent delivery, because the alternative is death. Surgical embolectomy is a last resort in specialist hands.

Secondary PPH. Resuscitate first, then treat the cause.

Resuscitate as for any obstetric haemorrhage. Assess consciousness, temperature, blood pressure, heart rate and respiratory rate; gain IV access; cross-match; and resuscitate exactly as for primary postpartum haemorrhage where the loss is significant — the principles of the structured haemorrhage response carry across (see postpartum-haemorrhage). Give a uterotonic: oxytocin 10 units IM as a single dose, then oxytocin 20 units in a litre of Ringer-lactate run at 125 mL/h, with ergometrine or syntometrine added unless the woman is hypertensive or has cardiac disease.

Endometritis-related bleeding — antibiotics and uterotonics. When endometritis is suspected, antimicrobials are started promptly; the evidence-supported regimen for postpartum endometritis is clindamycin plus gentamicin intravenously, and once uncomplicated endometritis has improved on IV therapy there is no benefit from a further oral course — a point worth holding, because the reflex to "complete a week of oral antibiotics" after clinical recovery is not supported by the trial evidence. The combination is chosen for its broad cover of the polymicrobial flora — anaerobes, Gram-negatives and streptococci — that ascend from the lower genital tract into the decidualised endometrium. In the SA district setting the NDoH regimen is amoxicillin plus metronidazole orally for mild puerperal sepsis, escalating to IV ampicillin, gentamicin and metronidazole given within an hour of presentation for severe sepsis, with transfer to specialist level; the antibiotic choices differ but the principle is identical — early, broad cover that treats the cause of the bleeding rather than just managing its consequence. Treating the infection often settles the bleeding because the involuting placental bed cannot heal while it is inflamed; antibiotics and uterotonics together address both the cause and the symptom.

Retained products — evacuate carefully. If products are retained and bleeding continues, surgical evacuation is the answer — but it carries a uterine-perforation rate of around 1.5% and a real risk of Asherman's syndrome, so it must be undertaken or supervised by an experienced clinician, ideally with ultrasound guidance, on a soft puerperal uterus that is easy to perforate. The puerperal uterus is thin-walled and softened, so the threshold for perforation is far lower than in a non-pregnant cavity; gentle technique, the smallest effective instrument, and a willingness to stop short of aggressive curettage all reduce the harm. Because uterotonics and antibiotics often settle bleeding that is not from a large volume of retained tissue, evacuation is reserved for the woman who genuinely has significant retained products and ongoing loss, not performed reflexively on the strength of an equivocal scan. Antibiotic cover and oxytocics accompany the procedure, and the evacuation is timed for daylight hours with senior support available, not undertaken alone at night.

Vascular causes — do NOT curette. This is the decision that separates a competent answer from a dangerous one. If colour-flow Doppler or the clinical pattern (torrential, intermittent bleeding, often post-caesarean, with an empty or near-empty cavity) suggests a pseudoaneurysm or AVM, the treatment is uterine artery embolisation by interventional radiology, not evacuation. A curette into a pseudoaneurysm ruptures the capsule and precipitates catastrophic haemorrhage; embolisation is definitive and uterus-sparing in a haemodynamically stable woman, and it preserves fertility in a way that surgery often cannot. The practical sequence is to stabilise the woman, get a colour-flow Doppler (or, where available, angiography) to confirm the lesion and its feeding vessel, and refer to interventional radiology — who occlude the bleeding vessel selectively. Transcatheter arterial embolisation and intrauterine balloon tamponade are both established for secondary PPH with ongoing bleeding; embolisation in particular is the specific answer for the vascular causes. The constraint in much of the South African public sector is access: interventional radiology is a tertiary, often metropolitan, resource, so the district-level discipline is to recognise the pattern, resist the curette, resuscitate, and arrange urgent transfer — and to have a frank plan for laparotomy and, if necessary, hysterectomy or internal iliac artery ligation if embolisation cannot be reached in time and the woman is exsanguinating.

When all else fails. Severe secondary haemorrhage unresponsive to the above may need examination under anaesthesia, and rarely laparotomy and hysterectomy, at specialist level.

Guidelines compared

The substantive divergence is the duration of postpartum thromboprophylaxis: the international standard is a fixed minimum (10 days or 6 weeks) tied to the risk profile, while the SA public-sector default of prophylaxis only "while in hospital" — where postnatal stays after vaginal delivery are short — leaves the highest-risk window, the first weeks at home, uncovered. The SAVE study found that even in women correctly identified as at-risk, adherence to duration was the weakest link.

The evidence & the controversy

The thromboprophylaxis debate is one of calibration, not principle. Everyone agrees that the puerperium is the peak-risk period and that LMWH prevents VTE; the argument is how widely to cast the net. The RCOG GTG 37a model is deliberately inclusive — caesarean section alone, or any two persisting risk factors, triggers ten days of LMWH — which medicalises a large proportion of postnatal women and, in a high-caesarean-rate, resource-limited service, is genuinely difficult to deliver and audit. The South African evidence is sobering: the SAVE cohort found that of women assessed as at-risk, only about four in five received any prophylaxis, and the duration prescribed was routinely too short, with the antenatal and post-discharge periods particularly neglected. The defensible position is not to dilute the risk assessment but to fix the delivery system — community LMWH prescription on discharge, taught self-administration, and a postnatal risk reassessment that does not stop at the hospital door — because the deaths are concentrated in exactly the women who were sent home without the prophylaxis their risk score demanded. Withholding ten days of enoxaparin from a class-3-obese woman after an emergency caesarean because "she's going home tomorrow" is the failure the confidential enquiries keep finding.

A second VTE controversy sits beneath the choice of imaging for suspected PE, and it is increasingly a topic of public and professional debate: the radiation trade-off between V/Q and CTPA. Neither is harmless, and the two harms fall on different people — V/Q's marginally higher childhood-cancer risk is borne by a hypothetical future child, CTPA's higher breast dose by the mother herself. Younger women with denser, more radiosensitive breast tissue may reasonably be steered towards V/Q where the chest X-ray is normal and local expertise allows, while an abnormal chest X-ray pushes towards CTPA because it answers the question a V/Q cannot. The defensible posture is to involve the woman in a decision that has no risk-free option, to use the chest X-ray to triage which test to do, and never to let the radiation anxiety delay imaging in a woman who is breathless and tachycardic — an undiagnosed massive PE is a far larger and more immediate risk than either scan.

The secondary-PPH controversy is a diagnostic one. The instinct that secondary bleeding equals retained products, and retained products mean a curette, is the source of avoidable disaster. The literature is consistent that endometritis, not retained products, is the commonest cause — in a high-caesarean-rate cohort it accounted for two-thirds of cases — and that ultrasound is too unreliable to mandate surgery on its own. The genuine clinical controversy is when to operate: an empty-looking cavity with ongoing bleeding does not need a curette, and a curette on a soft puerperal uterus harboring a pseudoaneurysm is lethal. The discipline that protects the patient is to add colour-flow Doppler to the assessment, to keep interventional radiology in the algorithm rather than reaching reflexively for evacuation, and to treat the infection that is usually the actual cause. The contemporary thread worth holding is the rising caesarean rate itself: as caesarean delivery climbs — globally and in SA — the aetiology of secondary PPH is shifting away from retained products and towards endometritis and the rare but deadly pseudoaneurysm, both of which are caesarean-associated. The disease is being reshaped by how we deliver.

Landmark trials & key evidence

The SAVE numbers are worth holding precisely: 126 of 220 women (57.2%) were at-risk, and only 104 of those 126 (82.5%) received any prophylaxis — so roughly one in five at-risk women received none at all, before counting those given too short a course. The intervention with the strongest mortality signal in this whole topic is not a drug but an audited, completed risk-assessment-to-prescription pathway.

Exam traps & red flags

• Curetting a pseudoaneurysm. Torrential, intermittent secondary bleeding — especially after caesarean, with a near-empty cavity — is a vascular cause until Doppler proves otherwise. The treatment is uterine artery embolisation; an evacuation ruptures the capsule and can exsanguinate the woman. This is the single most dangerous wrong answer in the topic.
• Using D-dimer to investigate VTE in pregnancy. It is physiologically elevated and uninterpretable peripartum. Diagnose with imaging; treat on clinical suspicion while you wait.
• Stopping at a negative early duplex. A negative compression duplex with high clinical suspicion of DVT means stop anticoagulation and repeat on days 3 and 7 — a propagating iliac thrombus can be missed on day 1.
• Under-dosing LMWH by weight. Defaulting to enoxaparin 40 mg in a 110 kg woman is under-prophylaxis of the highest-risk patient. Dose by weight band.
• Prophylaxis only "while in hospital". The peak-risk window is the first weeks at home. A high-risk or post-caesarean woman needs at least 10 days (6 weeks if previous VTE), prescribed for the community — not stopped at discharge.
• Giving a DOAC to a breastfeeding mother. Warfarin and LMWH are breastfeeding-safe; rivaroxaban and apixaban are not established in lactation and are avoided. Do not reach for a DOAC for convenience postnatally if she is breastfeeding.
• Treating secondary PPH as retained products by default. Endometritis is the commonest cause. Treat the infection, image sceptically (USS for retained products is unreliable), and do not operate on ultrasound alone.
• Forgetting βhCG. Persistent or recurrent secondary bleeding, particularly after a molar or non-viable pregnancy, needs a βhCG to exclude gestational trophoblastic disease.
• Missing massive PE in the collapsed puerpera. The shocked, breathless postnatal woman needs IV unfractionated heparin, urgent CTPA, and consideration of thrombolysis despite recent delivery — hesitation here is fatal.

Evidence anchors

• RCOG Green-top Guideline No. 37a — Reducing the Risk of Venous Thromboembolism during Pregnancy and the Puerperium (2015)
• RCOG Green-top Guideline No. 37b — Thrombosis and Embolism during Pregnancy and the Puerperium: Acute Management (2015)
• RCOG Green-top Guideline No. 52 — Prevention and Management of Postpartum Haemorrhage, BJOG (2016)
• Chainarong et al. — Secondary postpartum haemorrhage: incidence, aetiologies and clinical courses, PLoS One (2022)
• Naidoo et al. — Assessment and management of VTE risk during pregnancy and the puerperium (SAVE): the South African cohort, SAMJ (2019)
• Bistervels et al. — Intermediate-dose vs low-dose LMWH in pregnant/post-partum women with prior VTE (Highlow study), Lancet (2022): fixed low-dose is the appropriate prophylactic dose
• South Africa NDoH — Guidelines for Maternity Care in South Africa, Fourth Edition (2016): post-caesarean thromboprophylaxis (sodium heparin 5000 u 12-hourly or enoxaparin 40 mg daily while in hospital); suspected DVT/PE treated with enoxaparin 1 mg/kg twice daily; secondary PPH resuscitation, evacuation of retained products, and oral amoxicillin + metronidazole for subinvolution; puerperal sepsis antibiotic regimens.