An Albumin, Neutrophil, and Lymphocyte-Related Risk Estimation Tool in Hospitalised Patients

The neutrophil-to-lymphocyte ratio (NLR) is commonly used as a prognostic indicator for microbiological and inflammatory conditions in clinical settings. However, the quotient to albumin levels, which is another nutritional and clinical predictor, may also have an interesting diagnostic and prognostic value. This study aimed to primarily investigate the predictive performances of the neutrophils to albumin and lymphocytes ratio (NALR) compared to the NLR in predicting poor outcomes during hospital admission, particularly the decomposition of respiratory, renal, liver, and circulatory systems, resulting in longer hospital stays or mortality.

An observational study was performed on a cohort of 270 hospitalised patients admitted to Rashid bin Al-Hussein Military Hospital during the period from October 2023 to early November 2023. The study specifically targeted adult patients (age >17 years) who had a minimum of 80% availability of their initial and follow-up data during admission. We dichotomised all eligible test patients into two groups: Group I, which represented better outcomes of interest, and Group II, which represented poorer outcomes of interest. Statistically, we conducted binary logistic, receiver operating, and sensitivity analyses to explore the predictive performances and indices for NALR and NLR. We also conducted chi-square and independent T analyses to uncover the distribution rates of the independent variables across Groups I and II. We considered a p-value of less than 0.05 as the level of significance.

Out of a total sample size of 270, 82 patients (30.37%) were allocated to Group I, and 188 patients (69.63%) were allocated to Group II. Males outnumbered females in this study by 184 (68.1%) to 86 (31.9%). Patients in the study had an average age of 58.08±10.02 years. The average hospitalisation took 13.71±6.38 days, significantly longer in Group II compared to Group I (15.43±6.76 days vs. 9.77±2.69 days, p-value<0.05). We found that the area under the receiver operating characteristic (ROC) curves was estimated at [0.808±0.031 (0.748-0.868), p-value=0.000] and [0.667±0.034 (0.601-0.733), p-value=0.000] for NALR and NLR, respectively. The optimal operating thresholds for NALR and NLR were 1.5 and 5.37, with sensitivities and specificities of 86.7% versus 73.4% and 70.73% versus 70.73%, respectively.

The proposed NALR showed superior predictive performance, sensitivity, and correlation compared to the parent NLR. Both tools can be used in clinical practice to prioritise clinical and pharmacotherapeutics for hospitalised patients based on unfavourable outcomes.