When Kidney Function Falls in Cirrhosis: A Bedside Guide to HRS-AKI

What to Hold in Mind at the Bedside

Think of HRS-AKI as a kidney that behaves as if the patient were bleeding, inside a body that is holding too much salt and water. The circulation feels underfilled because the splanchnic arteries have dilated, so the kidney clamps down on its own blood flow even while ascites accumulates. That single idea explains why the same patient can be fluid-overloaded and kidney-underperfused, why a fluid bolus helps one patient and drowns another, and why the treatment that works is a vasoconstrictor that restores arterial filling rather than more volume.^1,13

• Recognize AKI from the change, not the absolute value. A rise of ≥0.3 mg/dL in 48 hours or ≥50% over the 3-month baseline is AKI, even when the number still looks normal. Sarcopenia, dilution, and bilirubin interference all make serum creatinine (SCr) understate the loss of filtration.^1,2
• Treat the precipitant and judge perfusion and congestion in parallel. When resuscitation is indicated, reassess SCr and urine output (UO) within 24 hours. When volume is already adequate or excessive, skip the fluid trial and move on.¹
• Apply all four ADQI–ICA elements — cirrhosis with ascites, AKI by SCr or UO, no improvement after indicated resuscitation, and no stronger alternative explanation.¹
• Once HRS-AKI is established, start a monitored vasoconstrictor promptly, individualize albumin to volume status, and start transplant evaluation in parallel — reversal is a bridge, not a cure.^1,3,4

Each major statement carries one of these labels. Verbs are calibrated: demonstrates and reduces are reserved for randomized evidence on that outcome; suggests marks single-trial or post hoc signals; is associated with marks observational relationships; may marks mechanism. Trial percentages are reported with their endpoint name, because the endpoints differ between trials and between the trial report and the product label.

Why Ascites and Poor Kidney Perfusion Coexist

Start with the normal kidney. Renal blood flow is held roughly constant across a wide band of arterial pressure by autoregulation: the afferent arteriole constricts when pressure rises and relaxes when it falls, so glomerular capillary pressure, and therefore filtration, stays stable. The kidney has its own reserve, and in health it seldom needs the rest of the circulation to defend it.

Now add portal hypertension. Scarring raises intrahepatic resistance, and portal pressure climbs. The splanchnic arterioles respond by dilating — driven by nitric oxide, carbon monoxide, endocannabinoids, and bacterial products that cross a leaky gut wall and stimulate local vasodilator release. Systemic vascular resistance falls. Even with a high cardiac output and an expanded total plasma volume, the arterial tree is relatively underfilled: the volume sits in the splanchnic bed rather than in the arteries that baroreceptors sense.

The body reads underfilling as hemorrhage. Baroreceptor unloading activates the renin–angiotensin–aldosterone system (RAAS), the sympathetic nervous system, and nonosmotic vasopressin release. Each response is appropriate for a bleeding patient and harmful here. Aldosterone and sympathetic tone drive avid sodium retention, which accumulates as ascites; vasopressin causes free-water retention and dilutional hyponatremia; and angiotensin II plus sympathetic tone constrict the renal vasculature. As cirrhosis advances, cardiac output can no longer rise to compensate — the reserve that some authors describe as cirrhotic cardiomyopathy — and the kidney loses perfusion pressure faster than it can autoregulate. That is HRS physiology: a kidney with intact structure, starved of effective perfusion.^1,12,13

Then add the modifiers. Infection — spontaneous bacterial peritonitis (SBP) above all — and systemic inflammation intensify vasodilation and can cause acute tubular injury (ATI) directly, which is why the ADQI–ICA panel no longer insists that HRS be "purely functional."¹ Large-volume paracentesis without albumin, bleeding, and overdiuresis abruptly reduce effective volume. Tense ascites can raise intra-abdominal and renal venous pressure, and right-heart dysfunction can add venous congestion; both lower the pressure gradient across the kidney from the venous side. Each modifier points to a different bedside decision, which the figure below links explicitly, often judged by the mean arterial pressure (MAP) response.

Why this matters for treatment. If the defect is arterial underfilling, the logical target is arterial filling. Splanchnic vasoconstrictors raise MAP and reduce splanchnic pooling; neurohumoral drive then falls and renal perfusion improves. In pooled patient-level data from the two largest terlipressin trials, terlipressin raised MAP by about 6 mmHg over placebo from day 1, and each 5-mmHg rise was associated with a higher hazard of reversal — a hazard ratio (HR) of 1.17 —, and MAP statistically mediated roughly one-third of the treatment effect.²⁰ That is a post hoc pharmacodynamic analysis, so it supports — rather than proves — MAP as a treatment target. Conversely, a patient whose MAP rises substantially on a vasoconstrictor while SCr does not fall probably has another process, most often tubular injury, and the ADQI–ICA panel advises re-evaluating for an alternative cause when MAP has risen ≥15 mmHg without SCr improvement.¹

Recognize AKI and Apply the ADQI–ICA Criteria

Choose the baseline deliberately. Use the lowest stable SCr from the previous 3 months; if none exists, the most recent value up to 12 months earlier may be used. When there is no prior value at all, the consensus suggests the lower of the admission SCr or a value back-calculated from an estimated glomerular filtration rate (eGFR) of 75 mL/min/1.73 m², and you should document that the baseline is uncertain.¹ Baseline choice is not paperwork: a falsely high baseline, taken after community-acquired injury began, hides the AKI and delays treatment.

Distrust the creatinine in both directions. Cirrhosis lowers creatinine generation through muscle loss, expands its volume of distribution through ascites and edema, and, in some assays, high bilirubin interferes with measurement; large-volume transfusion can lower it further. A creatinine of 1.3 mg/dL in a sarcopenic patient can represent a substantial fall in glomerular filtration rate (GFR). Standard eGFR equations assume a steady state and overestimate true GFR in cirrhosis, particularly with ascites or low GFR. Cystatin C is less dependent on muscle mass, can rise before creatinine, and adds prognostic information; use it where available, recognizing that it is not part of the diagnostic criteria.¹

Measure urine output properly. Hourly or timed UO is a sensitive early signal and oliguria predicts worse outcomes in critically ill patients with cirrhosis. Its value depends on the measurement: confirm catheter patency or collection reliability, record the method, and interpret UO in light of recent diuretic doses, which can produce urine without implying recovery.¹

What changed, and why it matters. The older International Club of Ascites criteria required 2 days of diuretic withdrawal plus albumin before HRS could be diagnosed. Uniform albumin loading caused harm through volume overload, and the wait delayed vasoconstrictors in a syndrome where response is better at lower creatinine.^1,13 The 2024 consensus therefore recommends against routine 48-hour albumin administration as a prerequisite for diagnosis, and anchors the decision to a 24-hour response only when fluid is genuinely indicated. Note also the UO wording: ADQI–ICA uses ≤0.5 mL/kg/h, whereas the generic Kidney Disease: Improving Global Outcomes (KDIGO) 2012 AKI definition uses <0.5 mL/kg/h.^1,2 Keep each wording attributed to its source. KDIGO's 2026 AKI and acute kidney disease (AKD) guideline update completed public review in May 2026; check its final status before relying on any revised staging language.

Terminology. Use HRS-AKI for the acute phenotype, and HRS-AKD or HRS-CKD for longer-duration hepatorenal dysfunction outside the acute window; these replace "type 1" and "type 2" HRS.¹ Describe superimposed disease explicitly — "HRS-AKI on CKD," "HRS-AKI with probable superimposed ATI" — because the treatment and the prognosis differ. Historical "type 1 HRS" remains relevant only because it defined eligibility in the pivotal trials, which used an SCr threshold of about 2.25–2.5 mg/dL; patients diagnosed today under the earlier 2024 criteria are, on average, less advanced than the trial populations.^1,4

Stabilize, Phenotype, and Reassess

The first shift has two jobs that run in parallel: treat anything that kills quickly, and assemble enough data to classify the AKI. Review baseline SCr, timed UO, MAP and perfusion, oxygen requirement, cumulative fluid balance, recent losses, bleeding, recent paracentesis, infection, and every drug given in the past week. Obtain urinalysis with microscopy, quantify proteinuria when the dipstick or history suggests it, and check for retention or obstruction with a bladder scan and kidney ultrasound when appropriate.^1,6

Do the diagnostic paracentesis. In a hospitalized patient with cirrhosis and ascites, prompt diagnostic paracentesis is part of the AKI work-up because SBP is the most common precipitant of HRS-AKI and is often clinically silent. An ascitic polymorphonuclear (PMN) count ≥250 cells/mm³ supports prompt SBP treatment. Send the cell count with differential and inoculate blood-culture bottles at the bedside, which improves culture yield.^1,6,11 Start empiric treatment for suspected shock or infection without waiting for the full AKI work-up.

Phenotypes overlap, and the tests are imperfect. A fractional excretion of sodium (FENa) below 1% is common in cirrhosis even without AKI, so a low FENa does not establish HRS; very low thresholds (below 0.1–0.2%) may be more specific but are often unmeasurable because laboratories do not report low urine sodium.¹ A bland sediment does not prove HRS either. Urinary NGAL is the best-studied injury marker: in a prospective U.S. cohort of 213 patients, a cut point of 244 µg/g creatinine separated acute tubular necrosis from prerenal AKI and HRS with a C statistic of 0.76, and higher NGAL was associated with 90-day death.⁹ Values overlap between phenotypes and depend on the assay, so treat NGAL as supportive information rather than a diagnostic switch. Pre-existing CKD, proteinuria, and tubular injury can all coexist with HRS-AKI; follow the trajectory and repeat the differential whenever the findings diverge from the expected course.

Match the fluid to the indication

Fluid choice follows the problem.¹ Use blood products for hemorrhage according to the bleeding pathway, and balanced crystalloid for ordinary volume depletion that requires resuscitation. When volume status remains equivocal and there is a reasonable prospect of fluid responsiveness, the ADQI–ICA management algorithm describes a limited challenge — 250–500 mL of crystalloid, or 20–25% albumin at 1–1.5 g/kg — followed by reassessment. Give a challenge only when the bedside assessment supports it, and use a combination of examination, imaging, and static or dynamic measures to judge the response, since no single method is superior.¹ If congestion or worsening oxygenation appears, stop further empiric fluid and reassess. Point-of-care ultrasound can help, but its signals (inferior vena cava size, venous Doppler patterns) are confounded by tense ascites and raised abdominal pressure, so read them with that limitation in mind.

Medicines. Stop nonsteroidal anti-inflammatory drugs (NSAIDs) and avoidable nephrotoxins, hold diuretics, and review kidney-cleared drugs. Nonselective beta-blockers reduce portal pressure and variceal bleeding risk, but in a hypotensive or AKI patient they can blunt the compensatory cardiac output the kidney depends on; reconsider them in the context of perfusion, bleeding risk, and congestion rather than stopping them reflexively.^1,6

Albumin by Indication, Then a Monitored Vasoconstrictor

Keep the albumin indications separate

Albumin is not one therapy. It appears in at least four distinct situations in cirrhosis, each with its own evidence base, dose, and ceiling, and confusing them is how patients receive several hundred grams in a few days. Review the cumulative albumin total whenever indications overlap.^1,5

Why albumin can help in one setting and harm in another. Albumin expands plasma volume and may bind inflammatory mediators, which plausibly explains its benefit in SBP, where acute vasodilation suddenly underfills the circulation. In a patient who is already euvolemic or congested — especially one who is also receiving a vasoconstrictor that raises afterload — the same volume moves into the lungs. The consistent message from ATTIRE and the terlipressin safety data is that albumin must be dosed against a reassessed volume state, never against a calendar or a laboratory target.^1,5,23

Choose the vasoconstrictor

Terlipressin is a vasopressin analog that acts mainly on V1 receptors in vascular smooth muscle. In the splanchnic bed, V1-mediated constriction reduces pooling and portal inflow, raises MAP, and unloads the neurohumoral response that is constricting the kidney. The same receptor sits on coronary, mesenteric, peripheral, and cutaneous vessels — which is where its ischemic adverse effects come from — and the rise in afterload with reflex slowing of the heart may contribute to pulmonary edema when it is combined with generous albumin.^1,23 ADQI–ICA names terlipressin the first-line vasoconstrictor, preferably as a continuous infusion, given promptly once HRS-AKI is diagnosed and paired with volume-guided albumin.¹ A small randomized trial found continuous infusion as effective as boluses at a lower daily dose, with fewer adverse events (35% versus 62%).¹⁵ The U.S. product label, however, approves bolus administration only — another reason to keep consensus and label statements apart.³

Norepinephrine is the monitored alternative when terlipressin is unavailable or unsuitable. Meta-analyses of small trials show similar HRS reversal with norepinephrine and terlipressin — for example, an odds ratio (OR) of 1.33 favoring terlipressin, with a 95% confidence interval (CI) of 0.80–2.22 that crosses no effect — with different adverse-effect profiles.^16,17 One open-label trial in ACLF found terlipressin superior (reversal 40% versus 16.7%; 28-day survival 48.3% versus 20%), so the equivalence may not hold in the sickest patients.¹⁴ Norepinephrine generally requires intensive care unit (ICU) monitoring and central or protocol-approved venous access; titrate it under the institutional infusion protocol, tracking MAP and kidney response. The consensus escalates norepinephrine when MAP has not risen by ≥10 mmHg and escalates terlipressin when SCr has not fallen by 25% from baseline.¹ Septic shock is treated through its own vasopressor and source-control pathway, not as HRS.

Midodrine plus octreotide has lower expected efficacy — in one meta-analysis, about 91% lower odds of reversal than terlipressin or norepinephrine.¹⁷ Reserve it as a fallback when terlipressin cannot be used and monitored norepinephrine is not feasible, and consider transfer when the required monitoring or organ support is unavailable.^1,6

Timing. Response is better when vasoconstrictors start at lower SCr — trial data suggest below about 2.25 mg/dL — which is a central reason ADQI–ICA removed the 48-hour wait.¹ In a small single-center randomized trial in ACLF (the eTerli study), starting terlipressin 12 hours rather than 48 hours after initial albumin increased day-7 AKI reversal (68.6% versus 31.4%).²¹ That result is consistent with the physiology but comes from one open-label trial and needs confirmation.

Monitoring on any vasoconstrictor. Follow SCr and UO alongside MAP, oxygenation, fluid balance, and signs of digital, mesenteric, or myocardial ischemia (abdominal pain, chest pain, cyanotic fingers, skin necrosis). If perfusion improves but kidney function does not, re-evaluate for ATI, infection, obstruction, glomerular disease, or ongoing congestion.¹

Response and stop rules — two sources, kept apart

Jurisdiction and product context

What the Trials Show — and What They Do Not

CONFIRM randomized 300 adults meeting historical type 1 HRS criteria to terlipressin or placebo, both with recommended albumin, for up to 14 days.⁴ Its primary end point, verified HRS reversal, required two SCr values ≤1.5 mg/dL at least 2 hours apart plus survival without KRT for at least 10 days after treatment. Terlipressin demonstrated a higher rate of verified reversal; it did not demonstrate a survival benefit, and death from respiratory disorders was more frequent.

The product label separately reports serious or fatal respiratory failure in 14% versus 5% — a different end point from the trial's respiratory death figure, so keep each number attached to its source.^3,4 Day-90 death was 51% versus 45%, a difference the trial was not powered to interpret. Two features limit extrapolation: the historical entry criteria enrolled patients with more advanced kidney dysfunction than today's earlier HRS-AKI definition captures, and patients with shock were excluded.⁴

Reading these results together. Vasoconstrictors consistently reverse HRS more often than placebo, and reversal is associated with less need for KRT before and after transplant: among CONFIRM patients who later received a liver transplant, those treated with terlipressin needed KRT less often at 12 months.²⁵ None of the trials has shown that a vasoconstrictor improves transplant-free survival, which is why the consensus frames treatment as a bridge to recovery or transplantation.^1,18 A subtler point: lowering SCr lowers the MELD score and can reduce transplant priority, which some allocation systems offset by holding the pre-treatment creatinine.¹

Transplant, Kidney Support, and the Months After

Liver transplantation is the definitive treatment for eligible patients, whether or not the kidney responds to a vasoconstrictor.¹ An episode of AKI in decompensated cirrhosis signals high short-term mortality, so start hepatology and transplant evaluation early — in parallel with nephrology and ICU management, and even if SCr improves. Refer to the transplant center with the data it needs: baseline and peak SCr, UO trend, duration of kidney dysfunction, KRT dates, and any evidence of underlying CKD. Liver-alone versus simultaneous liver–kidney (SLK) transplantation belongs with the transplant center and its criteria; in the U.S. allocation policy, sustained AKI (for example, eGFR ≤25 mL/min for at least 6 weeks, with or without dialysis), established CKD, and a post-transplant kidney "safety net" shape that decision.^1,26 MELD 3.0, which adds sex and albumin to the earlier sodium-adjusted model, is the current U.S. prioritization score.²⁷

Kidney replacement therapy is started for clinical indications — refractory hyperkalemia, acidosis, pulmonary edema, or uremic complications — not for a creatinine number.^1,2 Choose the modality by hemodynamics, local capacity, and team expertise; continuous KRT generally suits hemodynamically unstable patients and those at risk of cerebral edema. For patients whose recovery or transplant candidacy is uncertain, discuss goals, likely outcomes, and whether a time-limited KRT trial has defined milestones. The October 2026 global ACLF consensus adds critical-care statements on organ support and transplant decision-making in this population.^7,8 See the continuous renal replacement therapy (CRRT) guide for prescription and anticoagulation, and the AKI-on-CKD guide when HRS physiology sits on reduced baseline reserve.

• Transplant: hepatology and transplant referral documented on the day HRS-AKI is diagnosed
• KRT: indications reviewed daily; modality, access, and goals-of-care plan recorded
• Palliative care: personalized evaluation offered for symptom burden and goals, independent of transplant status
• Prevention of recurrence: SBP prophylaxis where indicated, albumin with large-volume paracentesis, NSAID avoidance, cautious diuretic restart
• Discharge summary: baseline and discharge SCr, residual dysfunction, suspected AKI mechanisms, medication changes, who rechecks labs and when

Follow-up. Almost half of patients with cirrhosis discharged after AKI are readmitted within 3 months.¹ Arrange a combined kidney and liver review within 1 month at minimum, and sooner for higher-risk patients; assess recovery and CKD status around 90 days; and screen for albuminuria, which identifies patients at higher risk of CKD progression. Creatinine overestimates kidney function after an illness that has wasted muscle, so use cystatin C where it is available.^1,2 The AKI after discharge guide covers the general 90-day pathway.

Where the Evidence Runs Out

Several questions that matter at the bedside remain unanswered. The pivotal trials used historical criteria, so response rates and harms under the 2024 definition — which captures patients earlier and at lower creatinine — are not directly known; better response is plausible but unproven.¹ The best MAP target, absolute or relative, has not been tested prospectively.^1,20 The optimal albumin dose during vasoconstrictor treatment is undefined, and the safety signal from ATTIRE and CONFIRM argues for less rather than more.^1,5,23 Head-to-head trials of terlipressin and norepinephrine using current criteria are small, and the ACLF data come largely from one center.^14,16 Injury biomarkers such as uNGAL have promising discrimination, but thresholds under the new criteria, and whether biomarker-guided treatment improves outcomes, are untested.^1,9 Finally, new agents remain investigational: a selective V1a partial agonist, OCE-205, showed numerically higher improvement than placebo in a small phase 2 trial that stopped early, without a statistically significant difference.²⁸

Three Synthetic Cases Through the Pathway

Original, synthetic adult vignettes for teaching. Open each case for the answer and rationale.

Eight Questions From the Bedside