ARISS — Albumin Replacement Therapy in Septic Shock

1. Publication

• Title: Albumin Replacement Therapy in Septic Shock: A Randomized Clinical Trial
• Acronym: ARISS
• Year & Journal: JAMA Network Open, published February 19, 2026
• Citation: Sakr Y, Nierhaus A, Schumacher U, et al; SepNet Critical Care Trials Group and ARISS investigators. Albumin Replacement Therapy in Septic Shock: A Randomized Clinical Trial. JAMA Netw Open. 2026;9(2):e2559297. doi:10.1001/jamanetworkopen.2025.59297

2. Context & Rationale

Background: Albumin is a natural colloid with oncotic, anti-inflammatory, antioxidant, and nitric-oxide-modulating properties that have supported its use in critical illness for decades despite conflicting trial evidence. The landmark SAFE trial found similar overall mortality between 4% albumin and saline, but a sepsis subgroup (n=1218) trended toward lower mortality with albumin. The subsequent ALBIOS trial randomized 1810 patients with sepsis/septic shock to 20% albumin (target ≥3.0 g/dL) versus crystalloid; overall there was no difference, but a post-hoc subgroup with septic shock at enrollment showed reduced relative risk of death with albumin.

Research Question/Hypothesis: In adults with septic shock, does albumin administration (started 6–24 hours after shock onset, titrated to maintain serum albumin ≥3.0 g/dL for up to 28 ICU days) reduce 90-day all-cause mortality compared with standard crystalloid-based fluid therapy?

Why This Matters: Septic shock carries roughly 40% mortality and affects hundreds of thousands of patients annually. Albumin is costly, and prior trials left its role in septic shock (as opposed to sepsis broadly) unresolved.

3. Design & Methods

• Study Type: Prospective, multicenter, controlled, parallel-group, open-label RCT
• Setting & Centers: 23 ICUs in Germany (SepNet Critical Care Trials Group), October 21, 2019 – May 2, 2022 (enrollment); study stopped prematurely for low enrollment (COVID-19 impact on ICU capacity and research activity)
• Population:
• Inclusion: Adults ≥18y; clinically possible/probable/confirmed infection; vasopressor requirement to maintain MAP ≥65 mmHg for ≥1h despite adequate volume therapy; serum lactate <18 mg/dL; septic shock onset <24h before inclusion (first albumin dose given 6–24h after shock onset)
• Exclusions: Life expectancy <28 days or palliative care <6 months; end-of-life decision; prior trial participation; hypersensitivity to albumin; conditions where albumin may be harmful (e.g., decompensated heart failure, traumatic brain injury) or specifically indicated (e.g., hepatorenal syndrome); pregnancy/lactation
• Intervention: 20% human albumin, 60-g loading dose over 2–3h (within 6–24h of shock onset, within 2h of randomization), then titrated by daily serum albumin (40g if 2.5–<3.0 g/dL; 60g if 2.0–<2.5 g/dL; 80g if <2.0 g/dL) for up to 28 ICU days
• Comparator: Standard-of-care crystalloid resuscitation; guideline-permitted albumin use allowed at treating physician's discretion for hemodynamic stabilization (48.6% of control patients received some albumin)
• Blinding: Open-label (treating physicians unblinded)
• Statistical Power & Follow-Up: Powered for 1412 patients (706/arm) to detect a 15% relative mortality reduction (50%→42.5%) at 80% power, two-tailed α=0.05; only 440 randomized before early termination. Modified ITT analysis (all randomized with ≥1 post-randomization observation). Follow-up to 90 days.

4. Key Results

Subgroup analyses: No significant differences in 90-day mortality by baseline lactate, SOFA, APACHE II, or SAPS II strata.

Safety: 267 AEs in 121/222 patients (54.5%) in the albumin group vs 202 AEs in 105/218 patients (48.1%) in controls — frequency and severity similar between groups; no signal of harm.

Target attainment: More than half of albumin-group patients never reached the target serum albumin ≥3.0 g/dL, despite a more liberal dosing protocol (0–80 g/day) than ALBIOS (0–60 g/day) — attributed to illness severity and dilutional effect of large resuscitation volumes.

5. Internal Validity Assessment

• Randomization & Allocation: Automated internet-based randomization, stratified by lactate (≤81.08 vs >81.08 mg/dL) and center — concealment adequate.
• Protocol Adherence & Separation: Biochemical separation was achieved (higher albumin levels in treatment arm), but heavily diluted by high control-arm crossover: 106/218 (48.6%) control patients received albumin per physician discretion, permitted by protocol/guidelines.
• Blinding & Detection Bias: Open-label; objective mortality endpoint limits detection bias, but performance bias (differential co-interventions) cannot be excluded.
• Missing Data: Primary outcome available for 210/222 (albumin) and 209/218 (control); 16 patients had consent issues, 1 excluded, 4 lost to follow-up.
• Overall Internal Validity Conclusion: Low-Moderate — premature termination (26% of planned sample) left the trial substantially underpowered; per-protocol and modified per-protocol sensitivity analyses were directionally consistent with the primary null result.

6. External Validity Assessment

• Population Representativeness: Single-country (Germany), 23-center trial; nearly half of admissions were post-surgical (emergency surgery 48.2%), which may limit generalizability to primarily medical septic shock populations.
• Practice Context: Reflects high-resource European ICU practice; albumin 20% formulation used (not the 4% used in SAFE), limiting cross-trial comparability.
• Overall External Validity Conclusion: Moderate for European surgical-ICU-heavy septic shock populations; less clear for medical-predominant or resource-limited settings.

7. Strengths & Limitations

Strengths:

• Multicenter, randomized design with concealed allocation
• Rigorous, CONSORT-compliant reporting
• Confirms albumin safety (no excess AEs)
• Liberal, biologically plausible dosing protocol targeting a physiological threshold

Limitations:

• Premature termination (440/1662 planned patients) — severely underpowered
• Open-label design
• Substantial control-arm albumin crossover (48.6%) diluting between-group separation
• Predominantly post-surgical cohort limits generalizability to medical sepsis
• More than half of treatment-arm patients never reached target albumin level

8. Interpretation & Practice Impact

• Clinical Implications: Does not support routine albumin administration targeting serum levels ≥3.0 g/dL in septic shock; current guidelines already restrict albumin to adjunctive use during large-volume crystalloid resuscitation, and this trial does not change that.
• Mechanistic Coherence: Failure to reach target albumin levels in most patients, plus heavy control-arm crossover, means the trial tested a diluted contrast rather than a clean albumin-vs-none comparison.
• Systems-Level Takeaway: Given albumin's cost and this trial's null, underpowered result, routine protocolized albumin-target order sets are not supported by this evidence.

9. Controversies & Subsequent Evidence

• Editorial Commentary: The authors themselves emphasize the result is "inconclusive" due to premature termination rather than a definitive negative finding.
• Guideline Integration: Findings are consistent with existing Surviving Sepsis Campaign guidance, which does not recommend routine albumin use beyond adjunctive resuscitation; ARISS does not provide grounds for a guideline change in either direction.

10. Summary & Executive Takeaway

Summary: ARISS randomized 440 (of a planned 1662) adults with septic shock across 23 German ICUs to albumin-targeted therapy (≥3.0 g/dL) versus standard crystalloid care. Ninety-day mortality was statistically similar (43.3% vs 45.9%; RR 0.94, 95% CI 0.76–1.17; P=.71). Albumin was safe but did not demonstrate a survival benefit; the trial was underpowered due to premature termination.

Overall Takeaway: ARISS neither confirms nor excludes a mortality benefit from targeted albumin replacement in septic shock — it adds a safe-but-null, underpowered data point to an already inconclusive evidence base, and does not justify routine albumin-targeting protocols outside current guideline-permitted adjunctive use.

11. Bibliography

• Finfer S, et al. A comparison of albumin and saline for fluid resuscitation in the ICU (SAFE). N Engl J Med. 2004;350(22):2247-2256.
• Caironi P, et al; ALBIOS Study Investigators. Albumin replacement in patients with severe sepsis or septic shock. N Engl J Med. 2014;370(15):1412-1421.
• Sakr Y, Bauer M, Nierhaus A, et al; SepNet Critical Care Trials Group. ARISS protocol. Trials. 2020;21(1):1002.