Which Observations from the Complete Blood Cell Count Predict Mortality for Hospitalized Patients?

Abel N Kho, MD, MS; Siu Hui, PhD; Joe G. Kesterson, BS; Clement J. McDonald, MD


Journal of Hospital Medicine. 2007;2(1):5-12. 

In This Article


Some investigators have incorporated selected CBC measures, for example, white blood cell count and hemoglobin/hematocrit, into multivariable models that predict mortality or rehospitalizations.[6,7,9,23] However, CBC reports can include a spectrum of more than 40 distinct counts and morphologic findings. Our study was the first to take into account all the different variables in the complete blood count and differential to determine elements that independently predict a high risk of mortality.

In addition to age and sex, our multivariable analysis of the 45 CBC variables found 13 independent predictors of mortality. Five were observations about white blood cells: absolute leukocytosis, high band form cell count, the presence of metamyelocytes, the presence of toxic granules, and absolute lymphocytosis. Eight were observations about red blood cells: high hematocrit, low hematocrit, high MCV and the presence of macrocytes, high red cell distribution width, the presence of NRBCs, the presence of burr cells, and the presence of sickle cells. Because controlling for severity of illness by Charlson comorbidity scores did not significantly change the model, the CBC abnormalities among the predictors of mortality did not simply reflect how sick the patients were. Including the 10 most common admission diagnoses did not significantly attenuate our reported odds ratios, suggesting the CBC predictors did not merely reflect the primary reason for admission. Interestingly, however, admission for chest pain did correlate with a greatly reduced risk of 30-day mortality, which may reflect the low threshold that physicians have for admitting patients with this complaint. Admission for acute but ill-defined cerebrovascular disease independently predicted a 2-fold increased risk of 30- day mortality.

What is the message to physicians from this analysis? Physicians commonly order CBCs and may rely on quick heuristics to sift through the myriad findings in CBC reports. Our analysis focuses physician attention on high-impact findings in the CBC. We assume that physicians already consider low hematocrit, high hematocrit (a sign of fluid loss and/or chronic hypoxia), high WBC count, high band cell count, and the presence of metamyeloctes (left shift) as important prognostic indicators. These abnormal findings are routinely mentioned at morning report and in a physician's notes.

Physicians, however, may not appreciate the importance of other CBC findings that our analysis found are predictive of mortality. Macrocytosis and a high RDW count (indicating an abnormally wide distribution of red blood cell size) have not previously been reported as predictors of mortality. And although other studies have suggested that bands are not predictors of mortality,[11] our study found they were an important prognostic indicator, with an OR =1.59, approaching those of leukocytosis and anemia.

The most impressive predictors of mortality were burr cells, NRBCs, and absolute lymphocytosis. The multivariate ORs of these 3, ranging from 2.8 to 3.2, were the highest of any CBC finding. In univariate analysis, the first 2 were associated with mortality rates 8 to 10 times higher than that of the average admitted patient. There are anecdotal reports in the literature of burr cells being associated with ominous prognoses[24,25,26] and more robust statistical analyses showing NRBCs to be associated with increased mortality.[14] Lymphocytosis has also been reported as a mortality risk in patients with trauma and emergency medical conditions.[15,16] Our analysis has shown that, indeed, all 3 of these findings are strong independent predictors of mortality.

The presence of sickle cells was also a strong predictor, but of decreased mortality. Patients with sickle cells in their smear had a risk of death one third that of patients without sickle cells. This does not indicate a protective effect. Rather, patients with sickle-cell disease typically are young and admitted for pain control and other non-life-threatening conditions. The presence of NRBCs in patients with sickle-cell disease appears to be intrinsic to the disease itself and did not have the same implications for mortality as it did for other patients in our study.

The overall logistic model including age, sex, and admission CBC variables had a respectable c statistic for predicting 30-day mortality of 0.80. This compares well with findings in other multivariable models. For example, the APACHE II score used to predict the mortality of hospitalized critical care patients has a c statistic that ranges from 0.78 to 0.86.[3,27,28] The APACHE score uses the worst value from the first 2 days after admission for some of its predictors so it cannot provide as early a warning as the admission CBC, and it requires collection of significantly more data. The inclusion of more CBC findings in the APACHE model might increase its predictive accuracy.

Our multivariate analysis was based on a very large number of patient samples using data collected through routine clinical care. However, our study has a number of limitations. The analysis was done at only a single institution, and the exact logistic regression model may not apply to other institutions that have different case mixes and laboratory procedures. Our institution's reported 30- day mortality rate of 3.4% was lower than the 4.6%- 11.9% reported in studies of patients admitted to general ward services,[29–31] but this may be accounted for by the lower-than-average Charlson comorbidity scores in our study population. Our risk adjustment by Charlson comorbidity scores may not be as precise as a risk adjustment tailored for our particular institution.[32] Our 30-day mortality rate was calculated using state death tapes, which means we would have missed patients who died outside the state, although we believe this rarely happens. We developed predictive equations on the basis of 30-day mortality, so we cannot comment on whether the CBC elements predict mortality beyond 30 days. We analyzed most variables as either high or low or as present or absent. Increasing degrees of abnormality may further increase the predictive power of some variables. Finally, the CBC is only one of many tests and clinical findings; it may be that some of these other variables would displace some CBC variables and/or improve the overall predictive power at the time the admission laboratory tests were performed. In this initial study, we have described the prognostic implication of the CBC across a wide range of diagnoses. Future work will focus on the predictive power of commonly gathered variables in more specific conditions (eg, low white blood cell count in sepsis).

Physicians generally have an intuitive ability to identify patients who are seriously ill and at high risk of dying[33] and adjust their diagnostic and therapeutic efforts accordingly. Our analysis highlights the value that certain observations in the CBC, notably burr cells, NRBCs, and absolute lymphocytosis, add to physicians' assessments of mortality risk. Even after adjustment for age, sex, comorbidities, common admission diagnoses, and other variables in the CBC, the presence of these findings predicted a 3-fold increase in 30-day mortality. Identifying the "red flags" within this ubiquitously performed test can make the difference in premature discharge or inappropriate triage of patients. Busy physicians can choose from a wide selection of ever-improving diagnostic tests, yet the workhorse CBC can serve as a simple and early identifier of patients with a poor prognosis.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: