Evaluate and manage acute and chronic renal disease in pregnancy

In one line

Renal disease in pregnancy splits cleanly into chronic kidney disease (CKD) — where the prognosis is set before conception by eGFR, proteinuria and blood pressure — and acute kidney injury (AKI), where the job is to name the obstetric cause and decide whether the kidney or the pregnancy is the thing to treat; the recurring trap is mistaking superimposed pre-eclampsia for either.

This chapter assumes the hypertension-in-pregnancy basics and the pre-eclampsia and HELLP groundwork; it concentrates on the discrimination between subtypes, the subtype-specific management, and the judgement calls, rather than the gestational renal physiology covered there.

Assessment

The central discrimination is pre-existing CKD versus pregnancy-acquired AKI versus superimposed pre-eclampsia, because each demands a different response and they masquerade as one another (all give hypertension, proteinuria and a rising creatinine). The hypertension-in-pregnancy basics and pre-eclampsia and HELLP groundwork underpin what follows.

• Interpret renal function against pregnancy physiology, not the non-pregnant lab. GFR rises ~40–50%, so creatinine falls: a "normal-looking" creatinine of 80–90 µmol/L in the third trimester may signal substantial impairment. eGFR equations (CKD-EPI, MDRD) are not validated in pregnancy — do not quote an eGFR number; track the absolute creatinine trend and measured/estimated creatinine clearance instead.
• Stage and stratify the CKD before anything else. The pre-pregnancy eGFR band, the degree of proteinuria, and the presence of hypertension or a systemic disease (lupus, diabetic nephropathy) are the three levers that set risk — this is the TOCOS framework (see trials table). CKD stages 4–5 (eGFR <30) carry roughly a 1 in 3 chance of needing dialysis during or within a year of pregnancy.
• Quantify proteinuria correctly. Use a spot protein:creatinine (uPCR) or albumin:creatinine ratio, not 24-hour collections. The pregnancy threshold for abnormal is >300 mg/24 h — double the non-pregnant value — because physiological proteinuria rises. Pre-existing proteinuria that worsens is hard to read late in pregnancy (see the pre-eclampsia overlap below).
• AKI work-up is cause-led. Stage by KDIGO criteria, then localise: pre-renal (hyperemesis, haemorrhage, sepsis), renal (pre-eclampsia/HELLP, the pregnancy thrombotic microangiopathies — TTP, atypical HUS, acute fatty liver), or post-renal (the gravid uterus, especially over a single or transplanted kidney). Pregnancy-related AKI in low-resource settings is dominated by pre-eclampsia, sepsis and haemorrhage — the SA reality, and largely preventable.
• Angiogenic markers help where the diagnosis is muddied. Where CKD or chronic hypertension obscures the diagnosis of superimposed pre-eclampsia, a low PlGF / high sFlt-1:PlGF ratio points to placental disease rather than progressing nephropathy — the ratio is markedly higher in pre-eclampsia than in CKD, though the rule-out cut-off may need lowering when renal function is impaired.

Classifying the subtypes

What matters is not "renal disease" but the subtype, because each one carries a different mechanism, a different trajectory, and a different management lever. Three axes place every patient.

Axis 1 — CKD, by aetiology, because the cause predicts the pregnancy behaviour. Not all CKD at a given eGFR behaves the same:

• Glomerular disease (lupus nephritis, IgA nephropathy, primary glomerulonephritis). Immune-driven, proteinuric, flare-prone. The mechanism that matters is that pregnancy is a relative immune-tolerant state that can mask or unmask activity, and the kidney can deteriorate from disease activity rather than from haemodynamics alone. Lupus nephritis is the archetype: the determinant of outcome is depth and duration of remission, not the label. Quiescent for ≥6 months on a pregnancy-compatible regimen behaves well; active or recently-active disease drives flare, superimposed pre-eclampsia and fetal loss.
• Diabetic nephropathy. Hyperfiltration injury on top of pre-existing nephron loss; proteinuria often climbs steeply in the third trimester and the BP and glycaemic burden compound the renal one. The lever is pre-conception optimisation (glycaemia, BP, ACE-inhibitor withdrawal converted to a pregnancy-safe agent).
• Reflux/obstructive and structural nephropathy, single kidney, polycystic disease. Lower flare risk; the dominant risks are hypertension, UTI/pyelonephritis (and in obstruction, post-renal decompensation), and — in ADPKD — a hypertension and pre-eclampsia excess.
• Hypertensive/vascular nephrosclerosis. The renal lesion is the BP; control is renoprotection.

Axis 2 — the timing and reversibility of the insult (CKD vs AKI vs acute-on-chronic). A woman with stage 3 CKD who develops HELLP has both a fixed baseline and a reversible superimposed insult, and the management has to address each. The trap is to attribute a step-change in creatinine to "her CKD progressing" when it is an acute, treatable obstetric event.

Axis 3 — for AKI, the anatomical lesion, because that is the management. Pre-renal (restore perfusion), intrinsic (treat the specific cause — and here the microangiopathy subtype is the whole game, below), or post-renal (decompress). In the SA septic-abortion and sepsis population, bilateral renal cortical necrosis is a distinct intrinsic subtype that does not recover — it leaves the woman dialysis-dependent. Its mechanism is microvascular thrombosis of the cortex from severe sepsis/haemorrhage/abruption, and it must be recognised as irreversible rather than treated in expectation of renal recovery.

The thrombotic-microangiopathy subtypes

Every pregnancy TMA presents with the same triad — microangiopathic haemolytic anaemia, thrombocytopenia and AKI — yet the treatments diverge completely, and the diagnosis is made on the pattern and the timing, not on a single confirmatory test (which usually returns too late to wait for). This is the discrimination on which everything else turns.

The most useful clinical rule: a microangiopathy that fails to improve — or that worsens — 48–72 h after delivery is not HELLP. That window re-sorts the diagnosis from "obstetric and self-limiting" to "primary TMA needing plasma exchange or complement blockade." The renal pattern helps before that: disproportionate, oligo-anuric AKI with prominent haemolysis but only modest hypertension and transaminitis argues against pre-eclampsia and toward aHUS or TTP.

Severity stratification and the subtle presentations

• The "normal" creatinine that is not. Because GFR rises, the woman with early CKD often books with a creatinine the lab flags as normal. Stratify against the expected gestational value, and treat a creatinine that fails to fall — or that creeps upward across trimesters — as the abnormal signal it is.
• The proteinuria you cannot interpret. A woman with baseline proteinuric CKD will, late in pregnancy, have rising protein from three possible sources: physiological gestational increase, her own nephropathy progressing, and superimposed pre-eclampsia. Proteinuria alone cannot separate these — which is exactly why the discrimination leans on BP step-change, new organ dysfunction, uric acid trend, and PlGF-based testing.
• The flare that hides behind pregnancy. In lupus nephritis, falling complement and rising anti-dsDNA can reflect either a flare or pregnancy's own immunological shifts; the discriminators are an active urinary sediment (cellular casts, dysmorphic red cells), a rising creatinine, and rising proteinuria with hypocomplementaemia — and the clinical context of how recently the disease was active.
• The post-renal lesion that masquerades as deterioration. A transplanted (pelvic) kidney or a solitary kidney can obstruct as the uterus grows; a disproportionately dilated system on ultrasound with a rising creatinine should prompt decompression (stent or nephrostomy), not escalation of medical therapy.

Management

Frame it immediate → ongoing → long-term, and run the maternal-renal and the obstetric threads in parallel.

Immediate. For AKI: resuscitate the cause — careful volume repletion (pre-eclamptic and septic kidneys flip to pulmonary oedema with over-filling, so titrate, don't flood), treat sepsis, relieve obstruction (ureteric stent/nephrostomy), and stop nephrotoxins. If a thrombotic microangiopathy is in play, the decision is mechanistic: HELLP/pre-eclampsia resolves with delivery; TTP needs plasma exchange; complement-mediated (atypical) HUS needs eculizumab — this triage carries the management. For pre-existing CKD presenting in pregnancy, the immediate task is the booking bundle: confirm safe medication, set the BP target, and start prophylaxis.

Staging AKI and the indications for dialysis

Stage the AKI explicitly with KDIGO — it disciplines the assessment and frames the urgency. Stage 1 = creatinine rise ≥26.5 µmol/L within 48 h, or 1.5–1.9× baseline, or urine output <0.5 mL/kg/h for 6–12 h; stage 2 = creatinine 2.0–2.9× baseline, or output <0.5 mL/kg/h for ≥12 h; stage 3 = creatinine ≥3× baseline or ≥353.6 µmol/L, or output <0.3 mL/kg/h for ≥24 h or anuria ≥12 h, or any need for renal replacement therapy. The crucial caveat: because the baseline creatinine in pregnancy is low, a woman can meet stage-2/3 criteria at an absolute value that looks unalarming on the non-pregnant reference range — stage on the fold-change from her own gestational baseline, not the lab flag.

The indications for dialysis are the standard ones — refractory hyperkalaemia, refractory acidosis, refractory fluid overload/pulmonary oedema, uraemic complications (encephalopathy, pericarditis) — but the threshold to start is lower in pregnancy, and the target differs from the non-pregnant patient: keep the pre-dialysis urea below ~12.5 mmol/L to protect the fetus from a urea-driven osmotic diuresis, polyhydramnios and the membrane stress of a high-urea milieu. Continuous modalities are gentler on placental perfusion than rapid intermittent dialysis when the woman is unstable.

Ongoing pharmacotherapy and targets.

Subtype-specific management

The generic table above is the floor; what changes for each subtype follows.

• Lupus nephritis. The non-negotiable is conceive in remission (≥6 months quiescent). Convert mycophenolate to azathioprine before conception; continue hydroxychloroquine throughout (it reduces flares, congenital heart block in anti-Ro/La–positive women, and pre-eclampsia, and stopping it is itself a flare trigger). A genuine renal flare in pregnancy is treated with corticosteroids and azathioprine, escalating to a calcineurin inhibitor if needed; cyclophosphamide is avoided in the first two trimesters and reserved for organ-threatening disease. The hard discrimination — flare versus superimposed pre-eclampsia — turns on the urinary sediment, complement/anti-dsDNA trend and PlGF: a flare has an active sediment and serological activity and may respond to immunosuppression; pre-eclampsia has the angiogenic signature and responds only to delivery. Getting this wrong means either immunosuppressing a placental disease or delivering a salvageable preterm fetus for a treatable flare.
• Diabetic nephropathy. Pre-conception ACE-inhibitor withdrawal (the renoprotection it gave non-pregnant is lost, and it is teratogenic), tight glycaemic and BP control, intensified aspirin prophylaxis, and counselling that third-trimester proteinuria will climb and does not by itself mean pre-eclampsia.
• Single/solitary or transplanted kidney with obstruction. Decompress rather than chase medically; a JJ stent or nephrostomy buys the pregnancy time and reverses a post-renal AKI that no amount of fluid or antihypertensive will fix.
• Renal cortical necrosis (the SA septic-abortion/sepsis subtype). Recognise it as a different prognosis: persistent anuria after the obstetric cause is controlled, with patchy non-enhancement on imaging, signals an irreversible cortical lesion — the management pivot is to establish maintenance dialysis and plan long-term renal replacement, not to keep waiting for recovery. Naming this early changes the counselling completely.

Transplant and dialysis. Counsel transplant recipients to wait ≥1 year post-transplant with stable graft function before conceiving; maintain tacrolimus, azathioprine and low-dose prednisolone (calcineurin-inhibitor levels often need up-titration as volume expands, so monitor trough levels through pregnancy), and switch mycophenolate to azathioprine ≥6 weeks pre-conception — mycophenolate causes a ~23% malformation rate (ear/facial/limb) and ~45–50% miscarriage. Across the pooled transplant data (Deshpande meta-analysis) the live-birth rate is ~73.5% with acute rejection in ~4.2% of pregnancies — so a successful pregnancy is the expectation, but acute rejection and graft loss are real, and the strongest predictors of postpartum graft loss are a pre-pregnancy creatinine above ~130 µmol/L and significant proteinuria. For women already on dialysis, intensify to long, frequent haemodialysis targeting a pre-dialysis urea <12.5 mmol/L (roughly ≥36 hours/week); the Hladunewich data show live-birth rate climbing with dialysis dose.

Surveillance and timing. Multidisciplinary care (obstetrician, nephrologist, neonatologist) with serial growth scans and uterine-artery Doppler for the predictable growth restriction (see fetal-growth-restriction), and serial creatinine/uPCR. There is no fixed delivery gestation — individualise on maternal renal trajectory, BP control, superimposed pre-eclampsia and fetal growth, generally not before 37 weeks unless one of those forces the hand. CKD is not itself an indication for caesarean.

Long-term. Counsel that pregnancy can accelerate decline: pre-pregnancy eGFR <40 with proteinuria >1 g/day predicts permanent loss of function, with transient deterioration in ~40% and permanent in ~30% of moderate-to-severe CKD. Arrange postnatal nephrology follow-up, breastfeeding-compatible drug review (methyldopa, enalapril and most CNIs are acceptable postpartum), and contraception that protects the kidneys and any future transplant (contraception-in-high-risk-women). For the woman whose AKI did not fully recover, name the acute-kidney-disease → CKD transition explicitly and book the nephrology follow-up that catches it — a single normal-looking discharge creatinine after a severe pregnancy AKI does not exclude residual nephron loss.

The evidence & the controversy

There is no safe CKD stage. The single most quotable shift came from TOCOS (Piccoli, JASN 2015): risk rises stepwise across stages, but even stage 1 CKD with no hypertension, no proteinuria and no systemic disease carried an OR of 1.88 for adverse outcome versus controls — there is a baseline risk intrinsic to the kidney disease itself. The meta-analytic backing (Al Khalaf, AJOG 2022) puts the adjusted OR for pre-eclampsia at 2.58, preterm birth 1.73, SGA 1.93 — the numbers that underwrite "high-risk pregnancy, multidisciplinary care." Translate the OR into counselling honestly: a pre-eclampsia OR of ~2.58 against a background risk that is itself elevated by hypertension and proteinuria means the absolute superimposed-pre-eclampsia risk in proteinuric stage 3–5 disease is substantial, which is what justifies aspirin, tight BP control and tertiary booking rather than a reassuring "most women are fine."

How tightly to control blood pressure is the live argument transplanted from the general hypertension literature. The renal guideline target (≤135/85) is tighter than CHIPS used; CHIPS (Magee, NEJM 2015) showed tight control did not worsen the perinatal composite while halving severe maternal hypertension, which underwrites aiming low in a population whose nephrons are already vulnerable — but CHIPS deliberately enrolled few women with significant renal disease, so the extrapolation is reasoned, not proven. In short: tight control is safe and protects the kidney; the trial evidence in CKD specifically is thin.

Dialysis dose changed practice without an RCT. The Toronto–US comparison (Hladunewich, JASN 2014) is observational, yet the dose-response — live-birth rates of 48% at ≤20 h/week rising to 85% at ≥37 h/week — is so steep and biologically coherent that intensive dialysis is now standard despite never being randomised; an RCT would arguably be unethical. The effect size is large enough to estimate the trade qualitatively: moving a woman from conventional (~20 h/week) to intensive (≥37 h/week) schedules roughly doubles the chance of a live birth — a difference no obstetric drug delivers, which is why the absence of randomisation does not weaken the recommendation.

Lupus nephritis outcome tracks remission depth, not the diagnosis. The PROMISSE data quantify it: renal flare in pregnancy occurred in roughly 8% of women in complete renal remission versus ~21% in partial remission — a near-tripling that operationalises the "conceive in remission, ≥6 months quiescent" rule and frames the answer when a woman with treated lupus nephritis asks whether she can safely start a pregnancy now. The corollary for counselling is to time the pregnancy to the remission rather than treat the pregnancy once it has flared.

Distinguishing superimposed pre-eclampsia from CKD progression is the hardest call in this group. In proteinuric CKD, new organ dysfunction or a step-change in BP after 20 weeks signals pre-eclampsia; PlGF-based testing (PARROT, Lancet 2019) halved the median time to clinical confirmation of pre-eclampsia and reduced severe maternal adverse outcomes — a genuine tool when proteinuria alone cannot tell nephropathy from placental disease. The unresolved edge is that the angiogenic cut-offs were derived in populations without significant renal disease, and CKD itself perturbs sFlt-1/PlGF clearance — so the rule-out threshold may need lowering in renal patients, and a borderline ratio in proteinuric CKD should be read cautiously rather than as a clean exclusion.

Tenofovir, the kidney and pregnancy — the SA-specific controversy. In the SA PVT and treatment population, tenofovir disoproxil fumarate (TDF) carries a small but real tubular and eGFR effect (on the order of a ~10% GFR decrement, occasionally proximal tubulopathy), whereas tenofovir alafenamide (TAF) delivers ~90% lower plasma tenofovir exposure and spares the tubule. Pregnancy-specific SA cohort data did not show a tenofovir-driven decline in renal function, so TDF is not withdrawn reflexively in pregnancy — but in a woman with established CKD or proteinuria, this is exactly where the regimen and the eGFR trend are reviewed with the HIV team, with renal function checked early (1–3 months) and then periodically.

Landmark trials & key evidence

Exam traps & red flags

• Quoting an eGFR in pregnancy. CKD-EPI/MDRD are not validated in pregnancy — track the absolute creatinine trend; a "normal" creatinine may be abnormal once the expected ~40% gestational fall is accounted for, and a woman can meet KDIGO stage 2–3 at an absolute value the non-pregnant reference range calls normal.
• Reading proteinuria with the non-pregnant threshold. Abnormal is >300 mg/24 h (or raised uPCR), and pre-existing proteinuria physiologically worsens — do not over-call pre-eclampsia on proteinuria alone; use BP step-change, organ dysfunction and PlGF.
• Treating the lupus flare as pre-eclampsia (or vice versa). Active urinary sediment + serological activity that responds to immunosuppression is a flare; the angiogenic signature that responds only to delivery is pre-eclampsia. Immunosuppressing a placental disease, or delivering a salvageable preterm fetus for a treatable flare, are opposite errors with the same root — failing to discriminate.
• Expecting recovery from cortical necrosis. Persistent anuria after the septic/haemorrhagic cause is controlled is renal cortical necrosis — an irreversible SA-context subtype — not slow-to-recover ATN; the pivot is to plan dialysis, not to keep waiting.
• Over-filling the pre-eclamptic or septic kidney — these patients drown in pulmonary oedema; titrate fluids, monitor output, and resist a reflex large bolus.
• Continuing an ACE-inhibitor/ARB into pregnancy (or mycophenolate in a transplant recipient) — fetal renal dysgenesis and major malformation; the switch happens pre-conception, not at the booking visit.
• Stopping hydroxychloroquine in lupus because "she's pregnant" — stopping it triggers flares; it is continued throughout.
• Missing a pregnancy thrombotic microangiopathy — TTP and atypical HUS mimic HELLP but do not resolve with delivery; a microangiopathy that worsens 48–72 h postpartum, or shows disproportionate oligo-anuric AKI with mild hypertension, needs plasma exchange (TTP) or eculizumab (aHUS), and failing to escalate is a "do-not-miss" emergency.
• Forgetting the SA context — pregnancy-related AKI here is dominated by pre-eclampsia, sepsis and haemorrhage, and HIV-associated nephropathy (APOL1-driven, proteinuric) is a real cause of underlying CKD; antiretroviral therapy as part of prevention of vertical transmission (PVT) is also renoprotective, and renal function needs monitoring on tenofovir.
• Defaulting to caesarean "because of the kidney" — route and timing are obstetric/maternal-trajectory decisions, not the renal diagnosis.

Worked viva — how to structure the answer

Consider a 31-year-old, 26 weeks, known lupus nephritis (last active 3 months ago), BP 158/102, creatinine 140 µmol/L (was 70 at booking), uPCR rising, platelets 130, an active urinary sediment. The management runs:

1. The clinical picture — this is recently-active lupus nephritis and acute-on-chronic kidney injury at 26 weeks; the first task is to separate a lupus flare from superimposed pre-eclampsia, because the treatments are opposite.
2. Discriminate — an active sediment, rising creatinine and proteinuria with serological activity point to a flare; check complement and anti-dsDNA and an angiogenic (PlGF / sFlt-1:PlGF) test — a placental signature re-points to pre-eclampsia. Stage the AKI by KDIGO against her gestational baseline, not the lab flag — a doubling from 70 to 140 is at least stage 2.
3. Stabilise and treat the right disease — control BP to ≤135/85 with a pregnancy-safe agent, continue hydroxychloroquine and aspirin, avoid nephrotoxins. If this is a flare, corticosteroids and azathioprine (escalating to a calcineurin inhibitor); if the angiogenic markers say pre-eclampsia, the lever is delivery, not immunosuppression.
4. Run it as a high-risk pregnancy — multidisciplinary (nephrology, obstetrics, neonatology), serial growth and Dopplers for FGR, serial creatinine/uPCR, and a clear escalation plan; book where neonatal ICU exists.
5. The evidence — TOCOS (no safe CKD stage), Al Khalaf (the OR numbers), PROMISSE (remission depth predicts flare), PARROT (PlGF to discriminate), CHIPS (the tighter BP target).
6. Close the loop — counsel that pregnancy can accelerate decline (eGFR <40 + proteinuria >1 g/day predicts permanent loss), arrange postnatal nephrology follow-up and the acute-kidney-disease→CKD watch, breastfeeding-compatible drugs, kidney-protective contraception, and — for the next pregnancy — conceive only once in sustained remission.

Evidence anchors

• Wiles K, et al. Clinical practice guideline on pregnancy and renal disease (BMC Nephrology 2019) — BP target, proteinuria, dialysis and transplant guidance
• TOCOS — Risk of Adverse Pregnancy Outcomes in Women with CKD (JASN 2015)
• Al Khalaf — CKD and adverse pregnancy outcomes: systematic review and meta-analysis (AJOG 2022)
• Hladunewich — Intensive haemodialysis and pregnancy outcomes (JASN 2014)
• CHIPS — Less-tight vs tight control of hypertension in pregnancy (NEJM 2015)
• CHAP — Treating mild chronic hypertension to <140/90 is safe and reduces adverse outcomes, no SGA excess (NEJM 2022)
• PARROT — PlGF testing in suspected pre-eclampsia (Lancet 2019)
• PROMISSE — kidney outcomes and predictors of nephritis in lupus pregnancy (CJASN 2017)
• Deshpande — pregnancy outcomes in kidney transplant recipients: meta-analysis (Am J Transplant 2011)
• Pregnancy-related AKI in the African continent: a systematic review (J Nephrol 2022)
• KDIGO Clinical Practice Guideline for Acute Kidney Injury (Kidney Int Suppl 2012)
• Husain — HIV-Associated Nephropathy in Africa: pathology, presentation, prevention (J Clin Med Res 2017)
• National Department of Health, Guidelines for Maternity Care in South Africa (knowledgehub.health.gov.za) — district→regional→tertiary referral and pregnancy-related AKI context