Routine blood measurement predicts in-hospital mortality
MedWire News: Measuring patients' red cell distribution width on hospital admission may help to gauge the risk that they will die during their hospital stay, say US researchers.
They draw their conclusion from an analysis of 74,784 patients admitted to the Beth Israel Deaconess Medical Center in Boston, Massachusetts. All patients had routine blood measurements on admission, including red cell distribution width, for which the center's normal reference range is 10.5-15.5%.
Overall, 1.3% of the patients died while in hospital. The likelihood for patients dying rose with increasing red cell distribution width, from 0.2% for those in the lowest decile (<12.9%) to 4.4% for those in the highest decile (>17.1%).
The association was strong and linear, say Sabina Hunziker and colleagues from the Beth Israel Deaconess Medical Center. Specifically, each 1% increase in patients' red cell distribution width raised their likelihood for in-hospital mortality by 24%, after accounting for multiple confounders.
The finding was consistent across subgroups including patients with myocardial infarction, liver disease, chronic renal failure, metastatic cancer, deficiency anemia, and alcohol abuse. It held true for men and women and for patients of different ethnicities.
Red cell distribution width also predicted admission to the intensive care unit; however, this association was less strong than that for mortality, at just a 4% risk increase per 1% rise in red cell distribution width after accounting for confounders. There was no clear linear relationship between red cell distribution width and need for intensive care.
"An interesting finding of this study was that red cell distribution width performed better for mortality prediction compared with prediction of acute deterioration of patients as reflected by intensive care unit transfer," say Hunziker et al in the American Journal of Medicine.
"Red cell distribution width is likely a marker for chronic processes leading ultimately to fatal outcome in inpatients and the general population, but seems less likely to reflect acute physiologic abnormalities, such as lactate or acute changes in blood gas analysis parameters."
The researchers also looked at other blood markers, but found that red cell distribution width predicted mortality better than did hemoglobin, hematocrit, platelet count, or mean corpuscular volume.
"After including these complete blood count values in the multivariate model, red cell distribution width was still a strong predictor for mortality, and discrimination of red cell distribution width was improved only slightly, suggesting that most prognostic information is covered by red cell distribution width," they conclude.
By Eleanor McDermid