QJM Advance Access originally published online on February 21, 2007
QJM 2007 100(4):217-223; doi:10.1093/qjmed/hcm006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Extreme leukocytosis in the emergency department
From the 1Department of Internal Medicine, Shaare Zedek Medical Center, Jerusalem, Israel 2Infectious Diseases Unit, and 3Clinical Hematology Laboratory
Address correspondence to Dr Y. Lawrence, Oncology Institute, Sheba Medical Center, Tel Hashomer, 52620 Israel. email: yaacovla{at}netvision.net.il
Received 27 December 2005 and in revised form 2 November 2006
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Background: Extreme leukocytosis in the absence of haematological disease, is a topic about which little is known, although it may be associated with increased mortality among patients admitted to the intensive care department. The significance of extreme leukocytosis in patients presenting to hospital is uncertain.
Aim: To study the correlates and prognostic significance of extreme leukocytosis, in patients admitted to an emergency department.
Design: Observational study.
Methods: Consecutive adult patients with extreme leukocytosis (>25 x 109/l, n = 54) presenting to the emergency department of a university-affiliated hospital were compared to age-matched controls (±5 years) with moderate leukocytosis (12–25 x 109/l, n = 118) presenting to the same department. Data were collected on demographic features, emergency room findings and hospital course.
Results: Patients with extreme leukocytosis were more likely to suffer from infectious disease (74% vs. 48%, p < 0.01), to be hospitalized (100% vs. 80%, p < 0.001), and to die (32.1% vs. 12.7%, p < 0.01), and had a longer median length of stay (7.5 vs. 4.0 days, p < 0.005). There was no significant difference in vital signs between the two groups.
Discussion: In our patients, extreme leukocytosis appeared to be predominantly caused by infectious disease, and was associated with a high case fatality rate. The degree of leukocytosis may provide prognostic information beyond that reflected in traditional vital signs.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Leukocytosis, an increase in the number of circulating white blood cells, was first described by Virchow1 and Andral2 in the mid 19th century. The leukemoid reaction, an extremely high white cell count resembling leukaemia, was first described by Krumbhaar in 1926.3 Early descriptions included many cases of what today would be classified as myeloproliferative diseases. There is little consensus regarding terminology. Previous authors5,8 have used the terms ‘leukemoid reaction’ and ‘extreme granulocytic leukocytosis‘, variously implying a leukocytosis in excess of 25–50 x 109/l in the absence of known haematological disease.
Today, the complete blood cell count is perhaps the most widely ordered medical investigation. However, despite the test's universal acceptance and long history, the prognostic and diagnostic significance of extreme leukocytosis in the modern clinical environment remains uncertain.
In moderate and extreme leukocytosis, the majority of cells are neutrophils.4,5 Neutrophilia may be caused by increased release from marrow stores, increased production, extended survival or demargination within the blood vessels. It is considered a non-specific measure of inflammation, being associated with bacterial and fungal infection, neoplasms, trauma, myocardial ischaemia, and with almost any medical condition that causes stress. Some medications including steroids, beta agonists and lithium may elevate the neutrophil count.6 Neutrophilia is also a feature of primary haematological disorders, including myeloproliferative disorders and haemolysis. Neutrophilia may be associated with the presence of immature neutrophils in the peripheral circulation (‘a left shift‘), including band forms, metamyelocytes and blasts.4
Previous studies have come to differing conclusions regarding the prognostic significance of an extremely high white-cell count. Reding et al.5 showed that adult patients who had had more than 25 x 109 leukocytes/l at any one point during their hospitalization had a 31% case fatality rate. However, no control group was presented for comparison. Waheed et al.7 found that patients with WBC >25 x 109/l at admission to an adult intensive care department had a higher case fatality than those whose WBC was 10–25 x 109/l. However, in two studies,8,9 extremely low-birth-weight infants with a very high white cell count had a slightly better prognosis than those with a lower white cell count. These studies are hampered by the lack of an adequate controls, heterogeneity in the patients studied, and small sample sizes.
Being unaware of any published study comparing moderate with extreme leukocytosis in the emergency room setting, we planned a study of consecutive patients presenting to the emergency department, comparing those with extreme leukocytosis (arbitrarily defined as leukocytes >25 x 109/l), with those with a moderate leukocytosis (defined as leukocytes 12–25 x 109/l).
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
This observational study was done at the Shaare Zedek Medical Center, a 550-bed university affiliated hospital serving the general Jerusalem population. The hospital contains all major services and departments except for neurosurgery and transplantations.
Patients aged >18 years who presented to the emergency department were included in the study if their initial leukocyte count was >25 x 109/l. Exclusion criteria were: known myeloproliferative or myelodysplastic disorders, or other haematological malignancy. The study ran from November 2003 to May 2004.
In the pilot study, for each patient-case with extreme leukocytosis, two non-sex-matched controls were chosen who conformed to the above criteria, except that their initial leukocyte count was between 12–25 x 109/l. Controls were selected as the next patients meeting the above criteria who presented to the emergency department on the same day as the study patient.
In the final study, two or three controls were chosen as above, except that they were age-matched controls (within ±5 years). If there were no appropriately aged patients on the same day, controls were selected from patients who presented the following day. Age-matched controls were likewise obtained retroactively for the pilot study patients.
Data were collected from the patients’ paper and electronic file, and, if necessary, from a brief interview with the patient. Parameters collected for each patient included demographic details, chronic diseases, medications, vital signs on admission, initial investigations performed in the emergency department, diagnosis made and outcome (admission, length of admission and death) as listed in Tables 1–4
. No additional tests were performed on patients for the purpose of this study. Data was collected within 48 h of admission.
|
|
|
|
Data were entered and analysed using the Epi Info 6.04d computer program (CDC). The statistical tests used were the
2 test, Fisher's exact test and two-tailed t-test, where applicable. Conditional logistic regression was used to compare the case and control groups after the data was stratified by age into 10-year blocks (<30 years, 30–39 years, 40–49 years, etc). Analyses used Stata version 9.2 (‘clogit’) (StataCorp). For assessing predictors of fatality in the pooled data, a binary logistic regression analysis was performed using the Epi 2000 version 3.2.2 program (CDC). A p value of 0.05 was set as significant. |
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
A pilot study of 30 patients with extreme leukocytosis and 59 controls with moderate leukocytosis was performed. Initial analysis indicated that the controls were significantly younger than the patients (60.0 ± 6.5 vs. 74.0 ± 6.4, p < 0.007). The protocol was therefore changed to obtain age-matched controls.
A total of 54 consecutive patients with extreme leukocytosis and 118 age-matched controls were recruited. The baseline characteristics, emergency department findings, investigations performed and outcomes are presented in Tables 1–4
.
No significant differences were found between the groups regarding demographic features or chronic diseases, except that more patients with extreme leukocytosis than controls were nursing-home residents (33% vs. 13.6%, respectively, p < 0.01) (Table 1). There were more patients who suffered from diabetes mellitus, and more use of steroids in the extreme leukocytosis group. Likewise the median age in the extreme leukocytosis group was 4 years older. However, none of these relationships reached statistical significance.
There was no difference in vital signs between the two groups (Table 2). Treating leukocytosis as a continuous variable, there was no correlation between leukocytosis and temperature at admission (correlation coefficient r2 = 0.01). Four patients with extreme leukocytosis, but only one with moderate leukocytosis were intubated at or before admission.
Table 3 shows that patients with extreme leukocytosis had a larger percentage of neutrophils in their blood, a more pronounced left shift (as signified by the greater percentage with excess band forms) and significantly more platelets. Patients with extreme leukocytosis were more likely to have a pathological chest X-ray. More patients with extreme leukocytosis had received prior antibiotic therapy, and they had a lower percentage of positive urine cultures.
Table 4 compares the diagnoses in the two groups. Infectious disease dominated in both, however this was more pronounced in the patients with extreme leukocytosis. The most frequent diagnoses in the extreme leukocytosis group were pneumonia (n = 16, 30%), urinary tract infection (n = 8, 15%), and Clostridium difficile infection (n = 5, 9%), compared to 16 (14%), 12 (10%) and zero, respectively, in the controls. No particular infectious agent dominated in the extreme leukocytosis group. Although most of the controls were hospitalized, more patients with extreme leukocytosis were hospitalized, predominantly in the internal medicine wing. Patients with extreme leukocytosis were more likely to be given antibiotics, had a longer length of stay, and had a higher case fatality rate than the control group.
Figure 1 shows the relationship between leukocytosis and case fatality rate. Bivariate analysis demonstrated a number of factors to be associated with in-patient death (Table 5).
|
|
Logistic regression analysis was performed on the variables which were associated with in-patient death at p < 0.05 in the bivariate analysis: age, degree of leukocytosis, mean arterial blood pressure, blood oxygen saturation measured by pulse oximetry, serum sodium, serum potassium, serum creatinine, and dwelling. Urinalysis was not included, since lack of data would have weakened the analysis. The result of blood cultures was excluded since they are not known at admission. Whether antibiotics had been previously given prior to admission was likewise excluded, since this is dependent on local policies, and would limit the generalizability of the results. Creatinine (p < 0.01, OR 3.9) was the only significant variable, with oxygen saturation (p = 0.05, OR 0.9) of borderline significance (Table 6). The degree of leukocytosis was not an independent prognostic factor for in-hospital death.
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
The discovery of the leukocyte is generally assigned to Hewson in the latter half of the eighteenth century, and subsequently, Virchow and Andral described leukocytosis. However, it was not until the 1870s that complete blood counts could be accurately performed, enabling Ehrlich to examine the role of leukopenia and leukocytosis in clinical medicine, and Krumbhaar to investigate leukemoid reactions.11
Today the crude blood count remains the cornerstone of clinical medicine. However, despite great advances in the understanding of the immune system made in recent decades, little progress has been made regarding the clinical significance of extreme leukocytosis.
We compared moderate and extreme leukocytosis in adult patients presenting to the emergency department. In the initial pilot study, the patients with extreme leukocytosis were significantly older than those with moderate leukocytosis, challenging the notion that the elderly are unable to mount a high leukocytosis. It is not clear whether this represents changes in the inflammatory response with age,10 or is merely a reflection of the more severe diseases and infections that afflict the geriatric population.
In our second, age-matched study, patients with extreme leukocytosis were more likely to live in nursing homes, and have an acute infection. They were more likely to be hospitalized, and had longer hospitalizations, and an increased case fatality rate.
Interestingly, there was little difference in vital signs (blood pressure, pulse, body temperature, oxygen saturation, respiratory rate) between patients with extreme and those with moderate leukocytosis, suggesting that the white cell count may provide useful prognostic information beyond that elicited by clinical examination.
By performing our study in an active university-affiliated hospital, we hoped to study the significance of extreme leukocytosis in a fairly typical emergency department. We did not, however, have any control over which patients were referred to, or presented to, the emergency department, and hence entered the study. The social, financial and health-related factors that affect the patient mix in an emergency department are likely to vary from country to country, limiting the generalizability of our results. Ideally an APACHE II score would have been calculated for each subject, but arterial blood gas studies were not always available. Despite running the study for >6 months, the number of cases was small, and the size of our emergency room limited our ability to obtain closely age-matched controls.
The only comparable study we know of involved patients admitted to the intensive care department,7 although it was not specified whether patients were admitted from the emergency department or from in-patient wards. Presumably many had already been hospitalized for a number of days, with a deteriorating condition necessitating their admission. The relevance of these findings in other clinical settings is unclear.
In our patients, extreme leukocytosis was associated with a higher case fatality rate. This seems somewhat surprising, considering the innate immune system's importance in acute infection. Extreme leukocytosis may be a surrogate marker for organ dysfunction. Serum catecholamine concentrations, one of the factors responsible for acute leukocytosis,12 rise in hypoxia,13 and in septic shock,14 especially when the latter is associated with acute renal failure.15 This hypothesis is supported by our multivariate analysis: the only independent predictors of in-patient death were raised creatinine level and low peripheral oxygen saturation (borderline significance), both measures of organ dysfunction, rather than leukocytosis. Extreme leukocytosis is an ominous sign, but presumably not the cause of a high case fatality rate.
Future studies should examine the clinical significance of extreme leukocytosis in other clinical settings, and investigate whether the degree of leukocytosis is correlated with serum cytokine levels.
In summary, extreme leukocytosis is associated with infection, longer hospitalization and a high case fatality rate. It appears to give information beyond that provided by vital signs, but is not an independent predictor of in-patient death.
|
Acknowledgements |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
We are grateful to Michael Huerta, MD for his critical reading of the paper, and helpful comments.
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
1. Virchow R. (1858) Die Cellularpathologie in ihrer Begrundung auf physiologische und pathologische Gewebelehre.Berlin August Hirschwald.
2. Andral G. (1843) Essai d’Hematologie Pathologique.Paris Fortin, Masson and Cie.
3. Krumbhaar EB. (1926) Leukamoid blood pictures in various clinical conditions. Am J Med Sci 172 519–33.[Web of Science]
4. Wenz B, Gennis P, Canova C, Burns ER. (1986) The clinical utility of the leukocyte differential in emergency medicine. Am J Clin Pathol 86 298–303.[Web of Science][Medline]
5. Reding MT, Hibbs JR, Morrison VA, Swaim WR, Filice GA. (1998) Diagnosis and outcome of 100 consecutive patients with extreme granulocytic leukocytosis. Am J Med 104 12–16.[CrossRef][Web of Science][Medline]
6. Holland SM and Gallin JI. (2004) Disorders of granulocytes and monocytes. In Kasper DL, Braunwald E, Fauci AS, Hauser SL, Longo DL, Jameson JL (Eds.). Harrison's Principles of Internal Medicine. 16th New York McGraw-Hill pp. 349–57.
7. Waheed U, Williams P, Brett S, Baldock G, Soni N. (2003) White cell count and intensive care unit outcome. Anaesthesia 58 180–2.[CrossRef][Web of Science][Medline]
8. Hsiao R and Omar SA. (2005) Outcome of extremely low birth weight infants with leukemoid reaction. Pediatrics 116 e43–51.
9. Rastogi S, Rastogi D, Sundaram R, Kulpa J, Parekh AJ. (1999) Leukemoid reaction in extremely low-birth-weight infants. Am J Perinatol 16 93–7.[Web of Science][Medline]
10. Ginaldi L, De Martinis M, D'Ostilio A, Marini L, Loreto MF, Martorelli V, Quaglino D. (1999) The immune system in the elderly: II. Specific cellular immunity. Immunol Res 20 109–15.[Web of Science][Medline]
11. Poulis S and Paterakis G. (2005) The dawn of blood—First seeing and then measuring. Haema 8 360–80.
12. Mac Gregor RR. (1977) Granulocyte adherence changes induced by hemodialysis, endotoxin, epinephrine, and glucocorticoids. Ann Intern Med 86 35–9.[Web of Science][Medline]
13. Nishikawa T and Naito H. (1996) Clonidine modulation of hemodynamic and catecholamine responses associated with hypoxia or hypercapnia in dogs. Anesthesiology 84 672–85.[CrossRef][Web of Science][Medline]
14. Benedict CR and Grahame-Smith DG. (1978) Plasma noradrenaline and adrenaline concentrations and dopamine-beta-hydroxylase activity in patients with shock due to septicaemia, trauma and haemorrhage. Q J Med 47 1–20.[Web of Science][Medline]
15. Wang W, Falk SA, Jittikanont S, Gengaro PE, Edelstein CL, Schrier RW. (2002) Protective effect of renal denervation on normotensive endotoxemia-induced acute renal failure in mice. Am J Physiol Renal Physiol 283 F583–7.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||