When Albumin Misleads: Serum Albumin’s Prognostic Reliability in Non‑Critical Illnesses – A Systematic Review

Ahmed Elawad¹, Asim Ahmed², Abla Almalik³, Nikos Papadopoulos⁴, Eleni Kosta⁵

Abstract

Abstract. Serum albumin (SA) is widely used as a prognostic marker because it reflects nutritional reserves, inflammation, and disease burden. Its low cost and routine measurement make it appealing for risk stratification, yet its clinical interpretation in stable, non‑critical illnesses remains uncertain. We systematically reviewed studies published up to February 2026 that evaluated SA as a predictor of outcomes in non‑ICU adult populations. Searches of open‑access databases and manual screening yielded 24 observational studies encompassing approximately 80 000 participants across cardiac, renal, respiratory, surgical, oncologic and geriatric settings. Most studies were prospective or retrospective cohorts. We extracted populations, albumin thresholds and adjusted effect sizes and synthesised findings narratively. Hypoalbuminaemia consistently predicted all‑cause or disease‑specific mortality in maintenance haemodialysis, heart failure with cardiac resynchronisation, chronic obstructive pulmonary disease, community‑acquired pneumonia, stable coronary artery disease and major surgery; hazard ratios for mortality generally doubled to quadrupled when SA fell below 3.5–4.0 g/dL[1][2]. Conversely, a small ischemic stroke cohort and certain chronic kidney disease analyses showed minimal prognostic value of SA alone[3][4], highlighting the confounding effect of systemic inflammation and fluid status. Composite indices such as the blood urea nitrogen–albumin ratio, hemoglobin‑albumin‑lymphocyte‑platelet (HALP) score, age‑to‑albumin index and neutrophil percentage‑to‑albumin ratio improved risk prediction across diseases. Overall, SA is a useful but non‑specific marker; clinicians should interpret it alongside inflammatory markers and clinical scores rather than rely on it as a standalone prognosticator.