Assess and counsel a high-risk patient requesting contraception, applying current eligibility criteria

In one line

For a high-risk woman, the contraceptive decision is a risk–risk trade-off: the absolute risk of the method must be weighed against the far larger risk of an unintended pregnancy in her specific condition, and the eligibility framework (WHO MEC / UKMEC) is the tool you use to argue it — never a substitute for clinical judgement when she carries multiple problems.

The method basics are covered in the Intermediate course: contraceptive modalities and postpartum contraception. The harder layer is judgement — the multimorbid woman, the framework's edges, the drug interactions, and the consultant call when two guidelines disagree.

Assessment

Frame every high-risk consult around four questions: how dangerous is pregnancy for her, how effective must the method be, which methods are her condition's contraindications, and what does she want? Asking them is routine; resolving them when the answers conflict is the harder consultant task.

• Quantify the pregnancy risk first. Pregnancy is the comparator the MEC categories are built against. In pulmonary hypertension, peripartum cardiomyopathy with residual dysfunction, severe valvular disease or Fontan circulation, maternal mortality is high enough that the most effective method (LARC or sterilisation) is the safety priority, not estrogen avoidance alone. The modified WHO cardiac classification (mWHO IV) names the conditions in which pregnancy is contraindicated — and those are precisely the women for whom a failed method is a catastrophe, so effectiveness, not estrogen-avoidance, is the dominant variable.
• History: the condition and its severity/control; thrombotic, cardiac, cerebrovascular, hepatic and oncological history; migraine phenotype (aura = focal neurology, not just severe headache); current drugs (enzyme-inducers, ART, bosentan, lamotrigine, anticoagulants); smoking; BMI; parity and future fertility wishes; HIV status and PrEP use; and any malabsorptive bariatric surgery (Roux-en-Y / biliopancreatic diversion), which undermines every oral hormonal method.
• Examination/investigation is condition-directed: BP (a true ≥160/100 is a hard CHC contraindication), BMI, and — for cardiac patients — current functional status and echo data, because peripartum cardiomyopathy is sub-categorised by recovery of ventricular function.
• Interpretation: a UKMEC/WHO MEC category 3 does not mean "no" — it means risks usually outweigh benefits and the method may still be the best available option after specialist input. Category 4 is an unacceptable risk. Two separate category-2 conditions are not additive into a 4; you re-appraise the woman as a whole.

The advanced judgement: where the framework runs out

The category tables are built for single conditions. The multimorbid woman breaks that assumption, and three judgement principles separate a consultant answer:

• Categories are not additive — but the highest one is binding. A woman who is both obese (BMI ≥35, CHC category 3) and a smoker over 35 (CHC category 3–4) is not a "category 6". You take the most restrictive relevant category as the floor, then re-appraise holistically: multiple concurrent category-2/3 cardiovascular risk factors push CHC firmly to category 4 because the biology (compounding arterial risk) is genuinely synergistic, whereas two unrelated category-2 conditions are not. Distinguish risk factors that share a mechanism (additive or supra-additive) from those that do not.
• Initiation versus continuation can diverge. Several conditions carry a lower category for continuing a method a woman already uses than for starting it — a new migraine-with-aura in an established CHC user, for instance, is a stop-and-switch, but the framework's continuation column exists because the act of stopping is itself not risk-free if it precipitates an unplanned pregnancy. Quote the column you are actually in.
• The condition's trajectory matters more than its label. "SLE" is not one MEC entry — it splits by antiphospholipid-antibody (aPL) status, disease activity and immunosuppression. aPL-positive (with or without clinical SLE) makes CHC category 4 regardless of how quiet the lupus is, because the thrombotic driver is the antibody, not the flare. Take the relevant sub-phenotype, not the diagnosis, to the table.

Management

Structure counselling as method now → review → definitive plan, anchored to the eligibility categories (1 = no restriction; 2 = advantages generally outweigh risks; 3 = risks usually outweigh advantages, needs expert judgement; 4 = unacceptable risk).

The high-risk conditions that bar combined hormonal contraception (CHC) almost all permit progestogen-only methods and the copper IUD. Estrogen drives the arterial- and venous-thrombotic excess; remove it and most doors reopen.

*Category numbers follow WHO MEC 6th ed (2025) / UKMEC 2025. Ranges reflect "initiation vs continuation" or severity sub-categories. *Copper-IUD insertion carries a vasovagal-syncope risk that is dangerous in severe pulmonary hypertension or single-ventricle physiology — insert with cardiac cover/sedation, or prefer the implant.

Subtype-specific management — the conditions the Final drills

• Cardiac disease: the etonogestrel implant is often the method of choice — highly effective, estrogen-free, no insertion-related haemodynamic risk. Progestogen-only methods carry no demonstrated thrombotic excess. Avoid the vasovagal hit of intrauterine insertion in the most fragile (Eisenmenger, severe PAH) unless done under monitoring with full resuscitation capability — a single vasovagal syncope can collapse a fixed-output circulation. Where definitive contraception is wanted in mWHO III–IV disease, laparoscopic tubal occlusion under general anaesthesia carries its own anaesthetic risk in these women; a Mirena or implant placed at the time of another planned procedure (or hysteroscopic/postpartum) often beats interval sterilisation. See cardiac-disease-in-pregnancy.
• VTE / thrombophilia / postpartum-haemorrhage survivors on anticoagulation: estrogen-free throughout. The LNG-IUS is additionally useful to control anticoagulation-related heavy menstrual bleeding — a genuine therapeutic two-for-one. A woman on a DOAC or warfarin with heavy menses should be offered the LNG-IUS as treatment, not merely tolerated as contraception.
• Severe/poorly controlled hypertension or established arterial disease (overlapping hypertension-in-pregnancy-antihypertensives and pre-eclampsia-early-onset-severe): CHC is category 4; DMPA is category 3 in established vascular disease (a small, possibly adverse lipid/vascular effect), so the implant, POP or LNG-IUS are preferred over DMPA here, and LARC plus counselling on the high pregnancy risk is the plan.
• polycystic-metabolic-ovarian-syndrome (previously PCOS) with obesity (BMI ≥35): CHC is category 3, not absolute — weigh metabolic/VTE risk against the non-contraceptive benefit (cycle control, endometrial protection); the LNG-IUS gives endometrial protection without the estrogen load and is the pragmatic default when BMI and metabolic risk make estrogen unattractive.
• Obesity and the malabsorption traps. Obesity per se does not blunt the implant, injectable or intrauterine methods — but it changes two things the framework hides. First, levonorgestrel emergency contraception loses efficacy with rising weight (signal of reduced effect from around 70–75 kg; ulipristal holds longer but is also weight-sensitive), so the copper IUD is the emergency method of choice in a heavier woman, and ulipristal is preferred over LNG orally. Second, after malabsorptive bariatric surgery the combined and progestogen-only pills drop to category 3 — counsel a non-oral method (implant, injectable, LNG-IUS, Cu-IUD).
• Women living with HIV (PVT framing): all hormonal methods and IUDs are now category 1 including in women at high risk of HIV — the precautionary DMPA caveat was removed in 2019 after ECHO. The real issue is drug interactions: efavirenz lowers implant/POP/CHC levels (MEC 2; DMPA stays 1), whereas dolutegravir — SA first-line — has no clinically significant interaction (all methods category 1). PrEP (oral TDF/FTC or the dapivirine ring) does not restrict any method. Always counsel dual protection (condoms for STI/HIV) and offer emergency contraception access.

The interaction layer the Final rewards

Enzyme induction is the commonest avoidable high-risk-contraception error, and the rule is mechanistic, not a memorised list:

• Enzyme-inducers (rifampicin/rifabutin, carbamazepine, phenytoin, phenobarbital, topiramate at higher doses, efavirenz, St John's wort) accelerate hepatic clearance of steroids that are absorbed and circulate systemically — so they undermine CHC, POP and the implant (UKMEC 3 for the implant with an enzyme-inducer; an extra method is advised). They do not meaningfully affect DMPA (high tissue depot), the Cu-IUD (no hormone) or the LNG-IUS (local endometrial action) — these are the safe choices on an inducer.
• Lamotrigine is the bidirectional interaction. CHC lowers lamotrigine levels (risk of breakthrough seizures on the pill, and toxicity in the pill-free week) — so the interaction runs the other way and is a reason to prefer an estrogen-free method in a woman on lamotrigine monotherapy.
• Bosentan (used in PAH) is itself an inducer and a teratogen — these women need a highly effective non-oral method and are exactly the women for whom pregnancy is most dangerous, so the implant or LNG-IUS, often dual-method, is the answer.

The evidence & the controversy

The ECHO trial (Lancet 2019) is the appraisal centrepiece. This open-label RCT randomised 7,829 women in eSwatini, Kenya, South Africa and Zambia to DMPA-IM, copper IUD, or LNG implant. HIV incidence was high (~3.8/100 woman-years) but did not differ significantly between methods: DMPA-IM vs Cu-IUD HR 1.04 (96% CI 0.82–1.33). The trial was designed to detect a ≥50% difference, so it cannot exclude a smaller (e.g. ~30%) effect — the honest appraisal is "no substantial difference, residual small effect not excluded, but the population-level benefit of access outweighs it." On that basis WHO reclassified all methods to category 1 for women at high HIV risk. The decisive limitations are the open-label design (no placebo) and the absence of a no-method arm; the comparison was between contraceptives, not against nothing.

CHC and thrombosis — argue with absolute numbers. Background VTE is ~2 per 10,000 woman-years; levonorgestrel-CHC roughly 5–7; drospirenone/third-generation CHC ~9–12. These are all dwarfed by pregnancy (~29 per 10,000) and the puerperium (~300–400 per 10,000) — the ASRM guideline quotes ~40–65 per 10,000 woman-years postpartum. So withholding any effective method from a thrombophilic woman who then conceives may increase her net thrombotic risk — the central paradox of high-risk contraception, and the reasoning that should drive the decision rather than category memorisation. The relative numbers reinforce it: MEGA put any-COC VTE at OR 5.0 vs non-use with levonorgestrel the lowest progestogen (OR 3.6), which is why a 2nd-generation LNG pill is the default if a combined method is justified at all.

The route does not rescue you from the estrogen. A recurrent candidate error is to offer the patch or ring as a "safer" combined method to a woman with a soft estrogen contraindication. The non-oral combined methods are not lower-risk for VTE: the transdermal patch delivers a higher total daily ethinylestradiol exposure than many pills, and its VTE risk is at least comparable to — and by some cohort estimates higher than — combined pills (transdermal patch VTE incidence around 9–10 per 10,000 exposure-years, in the range of third-generation pills; nested case-control OR ~1.1 vs a norgestimate-35 pill). VTE risk tracks the estrogen, not the delivery route — so the patch and ring inherit every CHC category-4 contraindication. They are alternatives for adherence, never for risk.

Thrombophilia is where the synergy bites. The category-4 status of CHC in a known thrombogenic mutation is not arbitrary: in COC users, mild thrombophilia (factor V Leiden) raises VTE ~6-fold and severe deficiency (antithrombin, protein C/S) ~7-fold over non-carrier non-users, and the factor-V-Leiden × COC interaction is markedly supra-additive (joint OR ~19 vs neither). Severe inherited thrombophilia is the hard stop; an isolated heterozygous FVL is a strong relative contraindication where an estrogen-free method is almost always the right answer. Routine pre-prescription thrombophilia screening is not recommended — it is neither cost-effective nor sensitive enough to change the estrogen-free default a good history already mandates. Screening a sister of a FVL carrier "to decide about the pill" is the classic trap: you would withhold estrogen on the family history alone, so the test cannot change management and a negative result falsely reassures (it does not exclude every thrombophilia).

Effectiveness is itself a safety intervention — quantify it. The high-risk argument is not only "which method is contraindicated" but "which method will actually prevent the dangerous pregnancy". In the Contraceptive CHOICE Project (Winner 2012), unintended-pregnancy failure was 0.27 per 100 participant-years with LARC versus 4.55 with the pill, patch or ring — an adjusted hazard ratio of 21.8 for the short-acting methods. That ~20-fold real-world gap is driven almost entirely by adherence, and it is the evidence base for offering LARC first to any woman in whom a pregnancy carries serious risk. In a woman with Eisenmenger syndrome the difference between a 0.3% and a 5% annual failure rate is the difference between a sound plan and a potentially fatal one — which is why effectiveness, not estrogen status, leads the counselling in mWHO III–IV disease.

Where the frameworks diverge: WHO MEC and UKMEC (now CoSRH, 2025) agree on nearly all category-4 conditions but differ in nuance on initiation-vs-continuation, on the post-partum/breastfeeding VTE window (CHC is category 4 in the first 6 weeks postpartum for any woman with an additional VTE risk factor, and the early window is handled slightly differently between frameworks), and on the new BMI handling. When SA practice diverges, follow the NDoH National Contraception Clinical Guidelines and SAMF availability — implants and DMPA are the workhorse LARC/injectable in the SA public sector, the copper IUD is freely available, ulipristal is not reliably stocked (so the Cu-IUD is the emergency method of choice when oral EC efficacy is in doubt), and the SA HIV Consolidated Guidelines (2026) frame interactions around dolutegravir-based ART.

Landmark trials & key evidence

The studies below underpin high-risk contraceptive practice; the numbers, not the conclusions alone, carry the argument.

Screening — what to do and what to skip

The advanced point here is negative: high-risk contraceptive care is mostly about not over-investigating.

• No routine thrombophilia screen before prescribing (above) — a good personal and family thrombotic history sets the estrogen-free default that a test would only confirm.
• No routine lipid or glucose profile before CHC in a young, asymptomatic woman; investigate where the history flags it (familial hyperlipidaemia, known diabetes with vascular complications — itself a CHC category 3–4).
• Blood pressure is the one mandatory measurement before any CHC — a documented, correctly-taken BP, because uncontrolled hypertension is the single most important, cheaply detectable arterial contraindication, and a normal BP is the precondition for an estrogen method.
• STI risk assessment before IUD insertion in a higher-risk woman (the small insertion-window PID excess), but a positive HIV status is not a contraindication to an IUD — that caution is obsolete.

Long-term, follow-up & counselling

• Counsel on what changes the category over time. A woman started on a method when low-risk can become high-risk — she turns 35 and still smokes (CHC → category 3–4), develops migraine-with-aura, becomes hypertensive, or is started on an enzyme-inducer or lamotrigine. Build the review trigger into the plan: "if you get a new visual aura with your headaches, or you start a new tablet, come back before your next pill."
• Document the shared decision in a category-3 case. When you prescribe against a category-3 condition because the alternative (pregnancy) is worse for her, the medicolegal protection is the documented risk–benefit conversation and the specialist input — write it.
• Definitive contraception is a long-term-management decision, not a default. In a woman with completed family and a mWHO III–IV cardiac lesion, counsel that a highly effective reversible method (implant/LNG-IUS) may be safer than the sterilisation procedure itself (anaesthetic/surgical risk in fragile physiology), and that her partner's vasectomy is the lowest-risk definitive option of all. Effectiveness, surgical risk and reversibility are weighed together.
• Always pair the method with emergency-contraception access and dual protection — the high-risk woman is exactly the one who cannot afford a method-failure pregnancy, and the heavier or post-bariatric woman needs to know the Cu-IUD is her most reliable rescue.

Worked viva — how to structure the answer

Examiners give a stem like "a 34-year-old, BMI 38, on lamotrigine for epilepsy, mother had a DVT on the pill, requests 'something reliable'." A high-scoring answer runs:

1. Frame it as a risk–risk problem — "She has three things that matter: obesity, a drug interaction, and a family thrombotic history. I'll weigh each method's risk against the risk of an unintended pregnancy for her."
2. Triage to the framework, then beyond it — "CHC is unattractive here: obesity is category 3, the family history pushes me toward estrogen-free anyway, and CHC would also lower her lamotrigine and risk breakthrough seizures. I would not screen for thrombophilia — it wouldn't change my plan."
3. Name the method and why — "I'd offer LARC first for effectiveness: the implant or LNG-IUS. The implant is an enzyme-inducer concern, but lamotrigine is not an inducer, so the implant is fine; the LNG-IUS is unaffected regardless. I'd avoid the oral methods given the interaction, and counsel that if she ever needs emergency contraception the copper IUD is most reliable at her weight."
4. Justify from evidence — CHOICE for the ~20-fold real-world effectiveness of LARC over the pill; MEGA/van Vlijmen for why estrogen plus a thrombotic history is the wrong combination; the lamotrigine interaction direction.
5. Close the loop — dual protection, a documented shared decision, and a review trigger ("come back if your seizures change or you start a new drug").

Exam traps & red flags

• Adding up category-2s into a 4. Wrong. Re-appraise the whole woman; specialist input, not arithmetic — but genuinely synergistic risk factors (multiple arterial risks) do compound to category 4.
• Treating category 3 as a refusal. It means expert judgement — often the best real-world option after counselling and documentation.
• Offering the patch or ring as a "low-estrogen" workaround. They inherit every CHC contraindication; VTE risk tracks the estrogen, not the route — the patch is, if anything, at the higher end.
• Forgetting the copper-IUD vasovagal risk in severe pulmonary hypertension / single-ventricle physiology — a syncopal insertion can be fatal; the implant is safer.
• Quoting the old DMPA–HIV caveat. Post-ECHO it is gone — DMPA is category 1 even in high-HIV-risk women.
• Missing the enzyme/ART interaction direction. Inducers undermine the implant and POP — but not DMPA, the Cu-IUD or the LNG-IUS; on dolutegravir there is no interaction. And CHC lowers lamotrigine — the interaction runs both ways.
• Migraine with aura = arterial-stroke flag. CHC is category 4 the moment aura is present, regardless of headache severity — adding CHC to migraine-with-aura carries an ischaemic-stroke OR ~6 vs neither risk factor.
• Screening for thrombophilia "to decide about the pill". It cannot change a decision already made by the history, and a negative test falsely reassures.
• Giving oral levonorgestrel EC to a heavier woman. Efficacy falls with weight; the copper IUD is the emergency method of choice, ulipristal second.
• Prescribing the oral pill after malabsorptive bariatric surgery. It drops to category 3 — use a non-oral method.
• The do-not-miss emergency: a woman with a high-mortality pregnancy condition (PAH, recent cardiomyopathy, mechanical valve on warfarin) leaving without effective contraception. Effectiveness is the safety intervention — default to LARC or sterilisation and document the counselling.

Evidence anchors

• WHO. Medical eligibility criteria for contraceptive use, 6th edition (2025) — categories 1–4; ART, PrEP and breastfeeding recommendations (DOI 10.2471/B09566).
• WHO. Revised recommendations on hormonal contraceptive use for women at high risk of HIV (29 Aug 2019) — all methods reclassified to category 1.
• Ahmed K, et al. HIV incidence among women using DMPA-IM, a copper IUD, or an LNG implant: the ECHO trial. Lancet 2019 — HR DMPA-IM vs Cu-IUD 1.04 (96% CI 0.82–1.33).
• ASRM Practice Committee. Combined hormonal contraception and the risk of VTE: a guideline (2016) — absolute VTE risk by progestin and vs pregnancy/postpartum (~40–65 per 10,000 wy postpartum).
• van Hylckama Vlieg A, et al. The venous thrombotic risk of oral contraceptives: the MEGA case-control study. BMJ 2009;339:b2921 — any-COC VTE OR 5.0 (4.2–5.8); levonorgestrel OR 3.6 (2.9–4.6).
• van Vlijmen EFW, et al. Combined oral contraceptives, thrombophilia and the risk of VTE: systematic review and meta-analysis. J Thromb Haemost 2016 — mild thrombophilia + COC RR 5.89; severe RR 7.15.
• Khialani D, et al. The joint effect of genetic risk factors and combined oral contraceptives on venous thrombosis risk. Br J Haematol 2020 — factor V Leiden × COC joint OR 19.3 (13.9–26.8) vs neither.
• Champaloux SW, et al. Combined hormonal contraceptives among women with migraines and risk of ischaemic stroke. Am J Obstet Gynecol 2017 — migraine with aura + CHC stroke OR 6.1 (3.1–12.1).
• Winner B, et al. Effectiveness of long-acting reversible contraception (Contraceptive CHOICE Project). N Engl J Med 2012;366:1998–2007 — LARC failure 0.27 vs pill/patch/ring 4.55 per 100 participant-yr; adjusted HR 21.8 (13.7–34.9).
• US MEC 2024 (CDC/MMWR). Category definitions and condition tables — wording of categories 1–4.
• FSRH/CoSRH. UK Medical Eligibility Criteria for Contraceptive Use (UKMEC 2025) — condition-by-method categories (full PDF access-restricted).
• South African NDoH. National Contraception Clinical Guidelines (2019) — SA method availability, HIV and ART context.