Systematic review, meta-analysis, and meta-regression of the association of serial leukocyte counts in critically ill patients and mortality - 17/03/26

Doi : 10.1016/j.aicoj.2026.100054

Nishkantha Arulkumaran ^a, Fiona Dewar ^a, Gareth Ambler ^b, Maria Del Pilar Arias Lopez ^c, Claudia Dos Santos ^d,^e, Jan J. De Waele ^f, Jeffrey Lipman ^g,^h, José-Artur Paiva ⁱ, Jean-Francois Timsit ^j,⁎

^a Bloomsbury Institute of Intensive Care Medicine, Division of Medicine, University College London, London, United Kingdom

^b Department of Statistical Science, University College London, London, United Kingdom

^c Paediatric Intensive Care Unit. Hospital de Niños Ricardo Gutierrez Buenos Aires, Argentina, Argentine Society of Intensive Care. Management, Quality and Data Committee Buenos Aires, Argentina

^d The Keenan Research Centre for Biomedical Science of St. Michael's Hospital, Toronto, ON, Canada

^e Institute of Medical Sciences and Interdepartmental Division of Critical Care, Faculty of Medicine, University of Toronto, Toronto, ON, Canada

^f Department of Intensive Care Medicine, Ghent University Hospital, Department of Internal Medicine and Pediatrics, Faculty of Medicine and Health Sciences, Ghent University, Ghent, Belgium

^g Jamieson Trauma Institute, Royal Brisbane and Women's Hospital, The University of Queensland

^h University of Montpellier, Division of Anesthesia, Critical Care, Emergency and Pain Medicine, Nîmes University Hospital, Nîmes, France

ⁱ Intensive Care Medicine Service, Unidade Local de Saúde São João, Alameda Professor Hernani Monteiro, 4200-319 Porto, Portugal. Faculty of Medicine, University of Porto, Portugal. Infection and Sepsis Research and Development Group, Portugal

^j Bichat Hospital, Medical and infectious diseases ICU (MI2), France University Paris-Cité, IAME, INSERM, F-75018 Paris, France

^⁎ Corresponding author.

connectez-vous ou créez un compte

Bienvenue sur EM-consulte, la référence des professionnels de santé.
Article gratuit.

Connectez-vous pour en bénéficier!

Sous presse. Manuscrit accepté. Disponible en ligne depuis le Tuesday 17 March 2026
Cet article a été publié dans un numéro de la revue, cliquez ici pour y accéder

Abstract

Background

Serial leukocyte counts have been investigated as a prognostic marker in critical illness. Our aim was to explore which immune cell type (total leukocytes, neutrophils, lymphocytes or neutrophil: lymphocyte ration (NLR)) best predicted mortality in critically ill patients. To address this, we conducted a systematic review, meta-analysis, and meta-regression of studies published within the past 10 years, archived on PubMed and Embase. Only studies reporting immune cell counts in critically ill patients on admission and within the subsequent week were included. Meta-regression was conducted to assess the effect of illness severity (reflected by study mortality rate) and underlying diagnosis (with ‘general ICU patients’ as a reference group) on the difference between immune cell counts between survivors and non-survivors. Meta-analysis and meta-regression were performed using Stata (version 19.5, StataCorp) and data presented as mean difference with (95% confidence intervals).

Results

Thirty studies including 26,441 reporting patient mortality were included. Neutrophil: lymphocyte ratio (NLR) best predicted mortality and was lower among survivors at on days 0-1 (-0.29 (-0.45 to -0.12), p = 0.001), 2-4 (-1.02 (-1.57 to -0.48), p < 0.001), and 5-7 (-0.69 (-1.00 to -0.38), p < 0.001). Lymphocyte count was higher at equivalent time points. Meta-regression of NLR suggest that heterogeneity between studies was partly explained by underlying illness severity on days 0-1 (mean difference decreases by 0.13 (0.05, 0.20) for a 10% increase in mortality, p = 0.001), but not at other time points. Underlying diagnosis did not explain heterogeneity between studies. We were unable to perform a meta-analysis investigating the association between leukocyte counts and secondary infections as only three studies reported this data.

Conclusion

A higher NLR count was associated with increased mortality throughout the first week of critical illness. Further work is required to define of cut- off values using an ordinal scale of lymphocyte count that accurately reflects immune dysfunction and risk of mortality.