Skip to main content

Foundational Trials

CONFIRM

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Terlipressin plus albumin for the treatment of type 1 hepatorenal syndrome
  • Acronym: CONFIRM
  • Year: 2021
  • Journal published in: The New England Journal of Medicine
  • Citation: Wong F, Pappas SC, Curry MP, et al. Terlipressin plus albumin for the treatment of type 1 hepatorenal syndrome. N Engl J Med. 2021;384(9):818-828.

Context & Rationale

  • Background
    • Type 1 hepatorenal syndrome (HRS-1) represents rapidly progressive renal dysfunction in advanced cirrhosis, with very high short-term mortality and limited disease-modifying options beyond liver transplantation.
    • Vasoconstrictor therapy combined with albumin had longstanding biological plausibility (splanchnic vasodilatation, reduced effective arterial volume) and was used in many settings, but robust confirmatory evidence with a clinically meaningful endpoint and contemporary trial conduct was needed.
    • Earlier studies and varied endpoints (biochemistry-only reversal vs harder outcomes) left uncertainty about the balance between renal benefit, transplant bridging, and clinically important harms (particularly respiratory complications) in an advanced, fragile population.
  • Research Question/Hypothesis
    • In adults with cirrhosis, ascites, and HRS-1, does terlipressin (added to albumin) increase the rate of verified reversal of HRS compared with placebo (with albumin), without unacceptable harm?
  • Why This Matters
    • Renal recovery can reduce the need for renal-replacement therapy (RRT), support candidacy for transplantation, and potentially improve peri-transplant outcomes in a group with severe multi-organ dysfunction.
    • Because adverse events (especially respiratory failure) could plausibly offset renal benefits, a high-quality, blinded, multicentre efficacy-and-safety dataset was essential to inform practice and guideline recommendations.

Design & Methods

  • Research Question: Whether terlipressin plus albumin improves verified reversal of HRS-1 versus placebo plus albumin in adults with cirrhosis and ascites meeting HRS-1 criteria.
  • Study Type: Phase 3, randomised, multicentre, double-blind, placebo-controlled trial (sites in the United States and Canada).
  • Population:
    • Adults with cirrhosis and ascites with HRS-1 (serum creatinine ≥2.25 mg/dL with a trajectory consistent with doubling within 2 weeks), evaluated after diuretic withdrawal and an albumin challenge.
    • Key exclusions included severe intrinsic renal disease, serum creatinine >7 mg/dL, shock, sepsis/uncontrolled infection or inadequate antibiotic duration, and recent large-volume paracentesis ≥4 L within 2 days (per protocol details).1
    • Baseline illness severity was high (mean MELD ~33; mean Child–Pugh ~10; baseline serum creatinine 3.5 ± 1.0 mg/dL in both groups).
  • Intervention:
    • Terlipressin 1 mg intravenously every 6 hours.
    • Protocolised dose escalation to 2 mg every 6 hours at day 4 if serum creatinine had not decreased by ≥30% from baseline; treatment for up to 14 days or until HRS reversal or discontinuation criteria were met.1
    • Concomitant intravenous albumin was recommended/used as standard care (in 165/199 [83%] in the terlipressin group; mean total dose 199.4 ± 146.8 g over median 5.0 days).
  • Comparison:
    • Placebo intravenously every 6 hours with matching escalation rules to preserve blinding.
    • Concomitant intravenous albumin used in 92/101 (91%); mean total dose 239.5 ± 183.6 g over median 5.5 days.
    • RRT and other supportive care delivered per local practice.
  • Blinding: Double-blind (participants, clinicians, and investigators); matching placebo with protocolised titration to maintain concealment.
  • Statistics: Power targeted detection of an increase in verified HRS reversal from 13.28% (placebo) to 28.36% (terlipressin) with 90% power at a two-sided 5% significance level; planned total sample size 300 with interim monitoring (O’Brien–Fleming-type boundaries per protocol); primary analysis in the intention-to-treat population using stratified testing; multiple imputation used for selected endpoints where prespecified.1
  • Follow-Up Period: Treatment up to 14 days; key renal endpoints assessed through day 14 and day 30; RRT, transplantation, and mortality tracked through day 90.

Key Results

This trial was not stopped early. A total of 300 patients were randomised (199 to terlipressin; 101 to placebo).

Outcome Terlipressin + albumin Placebo + albumin Effect p value / 95% CI Notes
Primary: verified reversal of HRS 63/199 (32%) 17/101 (17%) Not reported P=0.006 Composite included serum creatinine ≤1.5 mg/dL twice (≥2 h apart) by day 14, no RRT for ≥10 days, and survival for ≥10 days.
Secondary: HRS reversal 77/199 (39%) 18/101 (18%) Not reported P<0.001 Renal endpoint without the full “verification” components.
Secondary: HRS reversal with no RRT through day 30 69/199 (35%) 18/101 (18%) Not reported P=0.001 Clinically relevant to avoidance of dialysis in advanced cirrhosis.
Secondary: received RRT by day 90 58/199 (29%) 39/101 (39%) Not reported Not reported Table 3 (time-point proportions reported; no P value provided).
Secondary: died by day 90 101/199 (51%) 45/101 (45%) Difference 6 percentage points 95% CI −6 to 18 No survival advantage despite higher HRS reversal.
Adverse event: respiratory failure 28/200 (14%) 5/99 (5%) Not reported Not reported Safety population denominators; higher respiratory toxicity signal.2
Adverse events leading to death (respiratory disorders) by day 90 22/199 (11%) 2/101 (2%) Not reported Not reported Likely contributor to the numerical mortality difference; also highlighted in the Supplementary Appendix.2
  • Verified reversal of HRS occurred in 63/199 (32%) with terlipressin versus 17/101 (17%) with placebo (P=0.006).
  • HRS reversal with no RRT through day 30 occurred in 69/199 (35%) with terlipressin versus 18/101 (18%) with placebo (P=0.001), with fewer patients receiving RRT by day 90 (29% vs 39%).
  • There was no improvement in 90-day survival (51% vs 45%), alongside a higher respiratory failure signal (14% vs 5%) and more deaths attributed to respiratory disorders (11% vs 2%).2

Internal Validity

  • Randomisation and allocation: Central randomisation with stratification and matching placebo supported allocation concealment and reduced selection bias.
  • Blinding and detection bias: Double-blinding with objective renal laboratory criteria for primary endpoint reduced detection bias; however, clinically overt adverse events (e.g., ischaemia, respiratory decompensation) could plausibly threaten blinding at bedside.
  • Dropout/exclusions: Intention-to-treat efficacy denominators were 199 vs 101 (total 300); safety denominators differed slightly (200 vs 99) because of treatment exposure/misdosing.
  • Protocol adherence: Albumin was administered in 165/199 (83%) vs 92/101 (91%), with mean total albumin exposure 199.4 ± 146.8 g (median 5.0 days) vs 239.5 ± 183.6 g (median 5.5 days).
  • Separation of the variable of interest: Terlipressin exposure was protocolised (1 mg IV q6h with escalation to 2 mg q6h at day 4 if creatinine response was inadequate), while placebo mimicked identical escalation rules.1
  • Baseline characteristics: Groups were broadly comparable in key severity markers (serum creatinine 3.5 ± 1.0 vs 3.5 ± 1.1 mg/dL; MELD 32.7 ± 6.6 vs 33.1 ± 6.2; mean arterial pressure 78.7 ± 12.1 vs 77.5 ± 9.4; SIRS 42% vs 48%).
  • Timing: Screening and randomisation occurred within a constrained window after meeting criteria (including albumin challenge), supporting early treatment relative to HRS-1 recognition; applicability to earlier HRS-AKI phenotypes remains uncertain.
  • Dose optimisation: Bolus q6h dosing with protocolised up-titration may not be the only viable strategy (continuous infusion strategies exist in the literature), and the trial design did not test alternative dosing approaches.
  • Outcome assessment: Primary endpoint was a composite integrating renal recovery plus short-term survival without RRT, increasing clinical meaningfulness but also potentially diluting interpretability versus single hard endpoints.
  • Statistical rigour: Prespecified power, stratified analysis, and the use of multiple imputation for selected endpoints were consistent with a regulatory-quality design; interpretation still hinges on composite endpoint components and competing events.

Conclusion on Internal Validity: Overall, internal validity appears moderate-to-strong, supported by double-blinding, objective endpoint definitions, and prespecified statistical methods, but tempered by composite endpoint complexity, imputation, and a clinically important harm signal that could influence adherence and co-interventions.

External Validity

  • Population representativeness: Participants had very advanced liver disease (mean MELD ~33; Child–Pugh ~10) and severe renal dysfunction (baseline creatinine ~3.5 mg/dL), aligning with sick inpatient populations but representing a late-stage phenotype of HRS compared with modern earlier HRS-AKI definitions.
  • Key exclusions: Exclusion of shock and uncontrolled infection limits direct translation to mixed ICU populations where sepsis, vasopressors, and multi-organ failure are common.
  • Health-system context: Conducted in US/Canadian centres with access to transplantation and RRT; applicability may be limited where these resources are constrained.
  • Comparators: Placebo-controlled design informs efficacy versus supportive care, but does not directly answer comparative effectiveness against ICU norepinephrine-based strategies used in many critical care environments.

Conclusion on External Validity: Findings are most generalisable to hospitalised patients with advanced cirrhosis meeting severe HRS-1 criteria in high-resource settings; generalisability is more limited for earlier HRS-AKI, septic/shocked ICU patients, and settings where close respiratory monitoring or transplant access is reduced.

Strengths & Limitations

  • Strengths:
    • Randomised, double-blind, placebo-controlled design with protocolised dosing and escalation.
    • Multicentre North American conduct and a predefined clinically anchored primary endpoint (renal recovery plus short-term survival without RRT).
    • Clear signal for improved renal endpoints (verified reversal and HRS reversal) with consistent direction across related measures.
  • Limitations:
    • No improvement in 90-day mortality (51% vs 45%) despite higher reversal rates, challenging assumptions that renal reversal translates into survival benefit.
    • Clinically important respiratory harm signal (respiratory failure 14% vs 5%; respiratory-disorder deaths 11% vs 2%), raising patient-selection and monitoring concerns.2
    • Use of a severe HRS-1 definition (high creatinine threshold and rapid progression) limits applicability to earlier HRS-AKI presentations.
    • Placebo comparator does not resolve comparative effectiveness versus norepinephrine-based ICU strategies.

Interpretation & Why It Matters

  • Renal efficacy is real
    CONFIRM demonstrated higher verified reversal (32% vs 17%) and HRS reversal (39% vs 18%), supporting terlipressin as an effective vasoconstrictor strategy for severe HRS-1 in appropriately selected patients.
  • Hard outcomes remain challenging
    Despite reduced use of RRT at multiple time points (e.g., 29% vs 39% by day 90), there was no survival signal at 90 days (51% vs 45%), reinforcing that advanced cirrhosis mortality is driven by multi-organ complications beyond renal vasodysregulation.
  • Safety constraints shape bedside use
    Respiratory failure and respiratory-related deaths were more frequent with terlipressin, implying that fluid status, albumin strategy, baseline respiratory reserve, and close monitoring are not optional adjuncts but core to safe implementation.2

Controversies & Subsequent Evidence

  • Benefit–harm tension (renal recovery vs respiratory toxicity): The accompanying editorial emphasised that terlipressin’s renal benefits must be interpreted alongside a meaningful respiratory harm signal, raising questions about patient selection, albumin/volume management, and how to operationalise safety monitoring in real-world practice.3
  • “Reversal without survival” debate: Correspondence highlighted that improved renal endpoints did not translate into increased 90-day survival or transplantation rates, and pressed for clarity on co-interventions (including albumin strategy and beta-blocker exposure) that might modify both benefit and harm; the investigators’ reply contextualised the endpoint choice and reiterated the complexity of mortality drivers in advanced cirrhosis.4
  • Post-CONFIRM synthesis: Subsequent meta-analytic work continues to support improved HRS reversal with vasoconstrictors, but survival effects remain inconsistent and safety (particularly respiratory complications) is emphasised as a key determinant of net clinical value.89
  • Guideline incorporation: Major post-2021 guidance and consensus documents addressing decompensated cirrhosis/HRS-AKI incorporate vasoconstrictor-plus-albumin approaches and explicitly acknowledge the need to balance renal efficacy with adverse-event risk, reflecting the CONFIRM-era evidence base.567
  • Risk stratification for pulmonary complications: Emerging observational work has explored pulmonary congestion/respiratory reserve as prognostic modifiers in HRS-AKI, aligning mechanistically with the respiratory adverse-event concerns raised by CONFIRM.10

Summary

  • CONFIRM was a multicentre, double-blind, randomised, placebo-controlled phase 3 trial of terlipressin plus albumin for HRS-1 (199 vs 101; total 300).
  • Verified reversal of HRS occurred in 32% with terlipressin versus 17% with placebo (P=0.006).
  • Terlipressin increased related renal endpoints, including HRS reversal (39% vs 18%; P<0.001) and HRS reversal with no RRT through day 30 (35% vs 18%; P=0.001).
  • There was no improvement in 90-day mortality (51% vs 45%; difference 6 percentage points, 95% CI −6 to 18).
  • Respiratory safety signals were clinically prominent (respiratory failure 14% vs 5%; respiratory-disorder deaths 11% vs 2%), strongly shaping real-world risk–benefit assessment.2

Overall Takeaway

CONFIRM established that terlipressin plus albumin improves clinically anchored renal recovery endpoints in severe HRS-1, including verified reversal and reduced RRT use, but it did not improve 90-day survival and carried a clear respiratory safety signal. Its landmark status rests on simultaneously clarifying efficacy and forcing the field to confront implementation-limiting toxicity, making patient selection and respiratory/volume monitoring central to modern practice.

Overall Summary

  • Higher verified HRS reversal with terlipressin (32% vs 17%; P=0.006).
  • Reduced RRT use at follow-up timepoints (e.g., day 90: 29% vs 39%).
  • No 90-day survival benefit (51% vs 45%; 95% CI for difference −6 to 18).
  • Respiratory harm signal (respiratory failure 14% vs 5%; respiratory-disorder deaths 11% vs 2%).

Bibliography