Nucleated red blood cells in the blood of medical intensive care patients indicate increased mortality risk: a prospective cohort study
© Stachon et al.; licensee BioMed Central Ltd. 2007
Received: 23 March 2007
Accepted: 5 June 2007
Published: 5 June 2007
In critically ill patients, the appearance of nucleated red blood cells (NRBCs) in blood is associated with a variety of severe diseases. Generally, when NRBCs are detected in the patients' blood, the prognosis is poor.
In a prospective study, the detection of NRBCs was used for a daily monitoring of 383 medical intensive care patients.
The incidence of NRBCs in medical intensive care patients was 17.5% (67/383). The mortality of NRBC-positive patients was 50.7% (34/67); this was significantly higher (p < 0.001) than the mortality of NRBC-negative patients (9.8%, 31/316). Mortality increased with increasing NRBC concentration. Seventy-eight point six percent of the patients with NRBCs of more than 200/μl died. The detection of NRBCs is highly predictive of death, the odds ratio after adjustment for other laboratory and clinical prognostic indicators being 1.987 (p < 0.01) for each increase in the NRBC category (0/μl, 1 to 100/μl, 101 to 200/μl, and more than 200/μl). Each step-up in the NRBC category increased the mortality risk as much as an increase in APACHE II (Acute Physiology and Chronic Health Evaluation II) score of approximately 4 points. The mortality of patients who were NRBC-positive on the day of relocation from the intensive care unit to a peripheral ward was 27.6% (8/27). This was significantly higher than the mortality of patients who were NRBC-negative on the relocation day (8.6%, 28/325; p < 0.01). On average, NRBCs were detected for the first time 14 days (median, 3 days) before death.
The routine analysis of NRBCs in blood is of high prognostic power with regard to mortality of critically ill patients. Therefore, this parameter may serve as a daily indicator of patients at high mortality risk. Furthermore, NRBC-positive intensive care patients should not be relocated to a normal ward but should obtain ongoing intensive care treatment.
Under normal conditions, the peripheral blood of healthy adults is generally free of nucleated red blood cells (NRBCs), which tend to be found in patients with severe diseases [1–5] who have a relatively poor prognosis [3, 4, 6–9]. In most of the earlier studies on NRBCs, the concentration was determined microscopically by a stained peripheral blood smear. With such a technique, it is difficult to detect NRBC concentrations of less than 200/μl . For several years, a more convenient and sensitive technique has been available in the form of mechanized blood analyzers. With such an analyzer, one can routinely determine NRBC concentrations of less than 100/μl [11–13] The results of our recent studies with this new technique indicate that the detection of NRBCs may serve as an early indicator in patients at increased risk of mortality: on average, the presence of NRBCs was detected 1 to 3 weeks before death [14, 15]. Furthermore, the analysis of the cytokine profile in the blood of NRBC-positive patients (without hematologic diseases) suggests that NRBCs may be considered a parameter that sums hypoxic and inflammatory injuries. This may be the reason why the appearance of NRBCs is a strong predictor of increased mortality [15–17].
Recently, we reported on the poor prognosis of surgical intensive care patients when NRBCs are found in the peripheral blood . In that study, for the first time, a systematic day-to-day screening for NRBCs in the blood of surgical intensive care patients was performed. Our study revealed that 32% of the surgical intensive care patients were NRBC-positive at least once. The detection of NRBCs was associated with a greatly increased mortality of 44% (versus 4% of NRBC-negative patients). The area under the curve amounted to 0.86. On average, NRBCs were detected nine days before death. Therefore, in the present study, we set out to establish whether the daily screening for NRBCs in medical intensive care patients could serve as an early indicator of medical intensive care patients at extremely high risk.
Our study revealed that 18% of the medical intensive care patients were NRBC-positive. The detection of NRBCs was associated with a greatly increased in-hospital mortality. More than 50% of the NRBC-positive patients died. Furthermore, the mortality was three times higher in patients who were NRBC-positive on the day of relocation from the intensive care unit to a peripheral ward compared to patients who were NRBC-negative on the relocation day. On average, NRBCs were detected 14 days before death. These results suggest that the routine daily measurement of NRBCs could aid in a daily risk assessment of medical intensive care patients.
Materials and methods
Subjects and protocol
All intensive care patients treated between April 2003 and January 2004 in the intensive care unit of the Department of Internal Medicine of Berufsgenossenschaftliche Universitaetsklinik Bergmannsheil GmbH (University Hospital, Ruhr-University Bochum, Germany) (n = 383) were included in this study. Patients younger than 18 years and patients after surgery were excluded from this study. To evaluate the prognostic significance of NRBCs in the peripheral blood of medical intensive care patients, we screened one blood sample of each patient each day by means of a Sysmex XE-2100 (Sysmex Europe GmbH, Norderstedt, Germany). Blood samples were routinely drawn in the morning. For statistical analysis, a patient was defined as NRBC-positive when NRBCs were detected in the blood at least once. Outcome was considered as in-hospital mortality. Ethical approval to conduct this study was granted by the Ethical Committee of Ruhr-University Bochum (reference no. 1982).
Blood count parameters (NRBCs, leukocytes, hemoglobin, and thrombocytes) were measured using a Sysmex XE-2100 blood analyzer in line with the manufacturer's recommendations. According to the manufacturer, the NRBC detection limit was greater than 19/μl. Stringent internal quality control measurements were performed, and the criteria of acceptance were fulfilled throughout.
Creatinine, alanine aminotransferase, and C-reactive protein were measured with an LX 20 analyzer (Beckman Coulter GmbH, Krefeld, Germany), and prothrombine time ratio was assayed with a BCS (Behring Coagulation System) (Dade Behring, Schwalbach, Germany), all in accordance with the recommendations of the manufacturers. The quality assurance of quantitative determinations was strictly performed according to the German Norm: Quality Assurance in Medical Laboratories (DIN [Deutsches Institut für Normung] 58936, 2000). The criteria of acceptance were fulfilled throughout. Retrospective analysis of the laboratory data revealed 0.3% missing values.
Data are presented as the mean ± standard error of mean. When samples were normally distributed, the differences between the data for survivors and deceased were analyzed using the t test procedure. When samples were not normally distributed, the Mann-Whitney test was used because this test does not require a normal distribution of data. In the case of categorical data, the Fisher exact test was used. Correlations were analyzed by Pearson or non-parametric Spearman correlation. A p value of less than 0.05 was considered statistically significant.
A receiver operating characteristic curve was obtained by plotting the true-positive proportion (sensitivity) against the false-negative proportion (1 – specificity). The area under the curve (C-statistics) was calculated by non-linear regression.
The prognostic significance of NRBCs and other risk indicators was assessed using multiple logistic regression. In this study, the logistic regression tries to estimate the relative effect that parameters have on the patients' outcome. This is facilitated by assuming a functional relationship (the 'logistic model') between variables and probability of outcome. Then, for all possible settings, every variable is given a relative importance that makes the actual observed event 'most likely', taking into account the effects of all other variables. This is called the 'maximum likelihood' estimate of the variables' influence. These coefficients provide a relative weighting for each variable. Moreover, they can be used to derive odds ratios for the variables. If the odds ratio differs significantly from 1, a significant prognostic power that is independent of the other variables considered may be assumed.
In a first step, laboratory data were analyzed with regard to mortality. If reasonable for calculation of the odds ratios, the data were categorized in up to four categories. That is, NRBCs were subdivided into four categories: 0/μl, 1 to 100/μl, 101 to 200/μl, and more than 200/μl. A backward selection multiple logistic regression analysis was performed by first including all parameters in a multivariate model and subsequently leaving out the parameters with the largest p values until no parameter with a p value greater than 0.25 was included. The calculations were carried out using SAS version 8.02 (SAS Institute Inc., Cary, NC, USA). The intention of this study was to evaluate the prognostic power of the presence of NRBCs in the blood with regard to the patients' in-hospital mortality risk.
Nucleated red blood cells and established risk models for intensive care patients
The prognostic significance of NRBCs was evaluated under consideration of established risk models: the Acute Physiology and Chronic Health Evaluation II (APACHE II) (first evaluated in 1985 ) and the Simplified Acute Physiology Score II (SAPS II) (first evaluated in 1993 ). The APACHE II severity index includes the following risk factors: body temperature, mean arterial pressure, heart rate, respiratory rate, blood oxygenation, arterial pH, sodium, potassium, creatinine, hematocrit, white blood cell count, Glasgow coma scale, age, and anamnestic data concerning severe organ insufficiency or immunocompromised states of health. The SAPS II considers the following risk factors: age, heart rate, systolic blood pressure, body temperature, blood oxygenation, urinary output, urea, white blood cell count, potassium, sodium, bicarbonate, bilirubin, Glasgow coma scale, chronic diseases (that is, malignancies and acquired immunodeficiency syndrome), and type of admission (that is, medical and unscheduled surgical). Both the APACHE II score and the SAPS II are determined from the most deranged (worst) physiologic value (for example, the lowest blood pressure or the highest white blood cell count) during the initial 24 hours after intensive care unit admission.
We included 383 medical intensive care patients. The mean age was 66.3 ± 0.8 years (range, 20 to 94 years). Two hundred twenty-five male (58.7%) and 158 female (41.3%) patients were included in this study. On average, patients were treated for 4.1 ± 0.3 days (n = 383) in the intensive care unit. Total mortality was 17.0% (65/383). The APACHE II score and the SAPS II amounted to 16.0 ± 0.5 and 35.2 ± 0.9, respectively.
Clinical data and main diagnosis of treatment of NRBC-positive (n = 67) and NRBC-negative (n = 316) patients
68 ± 2
66 ± 1
Intensive care treatment, days
8.7 ± 1.2
3.2 ± 0.2
Body mass index
27.8 ± 0.9
26.5 ± 0.3
APACHE II score points
SAPS II points
Acute coronary syndrome/AMI
16.4% (n = 11)
38.0% (n = 120)
10.4% (n = 7)
14.2% (n = 45)
11.9% (n = 8)
6.6% (n = 21)
23.9% (n = 16)
8.9% (n = 28)
16.4% (n = 11)
7.9% (n = 25)
6.0% (n = 4)
10.1% (n = 32)
4.5% (n = 3)
4.1% (n = 13)
1.5% (n = 1)
4.4% (n = 14)
9.0% (n = 6)
5.7% (n = 18)
Prognostic significance of nucleated red blood cells
Furthermore, the mortality of patients who were NRBC-positive on the day of relocation from the intensive care unit to a peripheral ward was 27.6% (8/27). This was significantly higher than the mortality of patients who were NRBC-negative on the relocation day (8.6%, 28/325; p < 0.01).
Overall, with regard to in-hospital mortality, NRBCs in blood showed sensitivity and specificity of 52.3% and 89.6%, respectively. The area under the curve was 0.72.
NRBCs were an early indicator of patients at increased mortality risk. On average, in NRBC-positive patients who died, NRBCs were detected for the first time 13.6 ± 3.8 days (median, 3 days; n = 34) before death.
After the first detection of NRBCs in blood and during the further course of intensive care treatment, when the NRBCs have disappeared from the circulation, the mortality again decreased. That is, when former NRBC-positive patients were again NRBC-negative for more than 4 days after the final detection of NRBCs in blood, the mortality decreased to 16.7% (1/6).
Nucleated red blood cells in relation to other clinical and laboratory risk indicators
Incidence of NRBCs in blood in medical intensive care patients in relation to the APACHE II and the SAPS II
Score range of risk model
Incidence of NRBCs
Spearman correlation of the nucleated red blood cell concentration with other laboratory parameters (n = 67)
Prothrombin time ratio
Multivariate odds ratio estimates of clinical and laboratory risk indicators for in-hospital mortality calculated by logistic regression (n = 383)
95% confidence limits
Leukocytes (> 10/nl)
Prothrombin time ratio (< 60%)
C-statistics for several risk indicators for in-hospital mortality of medical intensive care patients
Area under curve
NRBC, highest value
Leukocytes, highest value
Prothrombin time ratio, lowest value
Alanine aminotransferase, highest value
C-reactive protein, highest value
SAPS II, on admission
APACHE II, on admission
APACHE II, on admission + NRBC, highest valuea
To our knowledge, this is the first study in which the detection of NRBCs in the peripheral blood was investigated with regard to its prognostic significance for the intensive care mortality of medical intensive care patients. In earlier studies, we and others have shown that the detection of NRBCs is associated with a relatively poor prognosis [3, 4, 6–8, 10, 14, 21, 22]. In most of those studies, the NRBC detection and quantification were based on the microscopic analysis of stained blood smears. This technique is time-consuming and only partly suitable for the detection and quantification of NRBC concentrations of less than 200/μl .
In this study, the NRBC concentration was screened with a mechanized blood analyzer of high sensitivity [11, 12, 23]. The present study revealed that approximately 18% of all medical intensive care patients were NRBC-positive at least once. Interestingly, in nearly half of the NRBC-positive patients, NRBCs were detected already on the admission day.
Our data confirmed the high prognostic power of the mechanized detection of NRBCs in blood in terms of mortality. The total in-hospital mortality of NRBC-positive patients of this study was 50.7%. Furthermore, as shown in earlier studies, the present data showed that the mortality increased with an increasing NRBC concentration [10, 24, 25] Approximately 80% of the patients with NRBC concentrations higher than 200/μl died.
Our study revealed that the daily screening for NRBCs can be used to estimate the patients' mortality risk. Not only did the predictive value for death increase with the concentration of NRBCs in the blood, but the prognosis improved when the NRBC concentration decreased. In particular, after the first detection of NRBCs in blood and during the further course of intensive care treatment, when the NRBCs have disappeared from the circulation for more than four days, the mortality again significantly decreased nearly to values of NRBC-negative patients .
In the present study, increased creatinine and leukocyte concentrations and a lower prothrombin time ratio were significantly correlated with increased NRBC concentrations. Although these findings suggest that NRBC-positive patients are more severely burdened than NRBC-negative patients, the detection of NRBCs is an independent predictor of poor outcome. To evaluate the independent attributable risk factor, a logistic regression considering NRBCs, age, gender, body mass index, APACHE II score, creatinine, hemoglobin, thrombocytes, leukocyte, alanine aminotransferase, C-reactive protein, and the prothrombin time ratio was performed. As a result, the independent prognostic power of NRBCs is underlined by an odds ratio of 1.987 for each stepwise increase in the NRBC category. That is, patients with NRBCs of more than 200/μl have a more than seven-fold higher risk to die than NRBC-negative patients. In recent studies, we have already demonstrated that the detection of NRBCs is a risk indicator that is independent of several other established risk indicators [10, 14, 16, 24].
Among the general severity of illness scoring systems for intensive care patients, APACHE II and SAPS II have become two of the most accepted and used [26–30]. However, the present data suggest that the APACHE II score could be significantly improved by adding up to 12 score points, considering the presence of NRBCs as an independent variable in this score, as suggested for the abbreviated burn severity index in patients with burns .
The analysis of the lifespan of the patients who died indicates that NRBCs in blood were found not just immediately before death. Moreover, our present study showed that the detection of NRBCs is often a relatively early phenomenon prior to death. In deceased patients, NRBCs were detected 14 days before death. Therefore, NRBCs would seem to be an early indicator of increased risk.
Finally, the underlying pathophysiology of NRBCs in blood is not fully understood. In our study, no association with only one of the various causes of patient death was found. However, some authors have claimed that hypoxemia [31, 32], acute and chronic anemia [33, 34], or severe infections [35, 36] are linked to the appearance of NRBCs in critically ill patients. In this context, we recently reported on the cytokine profile and the erythropoietin concentrations in NRBC-positive patients . Our data suggested an important role of inflammation and/or decreased tissue oxygenation (caused by local or systemic circulatory disorders) for the appearance of NRBCs in blood. NRBCs may thus be considered a marker that sums up hypoxic and inflammatory injuries. It seems obvious that these complications have an impact on patient prognosis. Therefore, this could be the reason why the appearance of NRBCs is a strong predictor of increased mortality.
However, concerning such an association between inflammation (with or without hypoxia) and NRBCs, it is attractive to speculate what kind of therapy could improve the poor prognosis of NRBC-positive patients. Currently, studies are under way in our university hospital to show whether intensifying the treatment of patients with NRBCs (that is, an earlier administration of antibiotics or an anti-inflammatory therapy) can reduce their mortality rate.
Nonetheless, we observed that the mortality was three times higher in patients who were NRBC-positive on the day of relocation from the intensive care unit to a normal ward compared to patients who were NRBC-negative on the relocation day. Consequently, it seems obvious that NRBC-positive patients should obtain ongoing intensive care treatment.
This is the first study in which the daily screening for NRBCs in the peripheral blood of patients in the medical intensive care unit was investigated with regard to its prognostic power for in-hospital mortality. The incidence of NRBC-positive patients was 18%. NRBC detection in critically ill patients was associated with significantly increased in-hospital mortality (50.7% versus 9.8%). The predictive value for death increased with the NRBC concentration and seems to decline again when the NRBCs have disappeared from the circulation. The prognostic significance of NRBCs was independent of other laboratory and clinical risk parameters. An improvement of established risk models like APACHE II seems feasible. Furthermore, the detection of NRBCs in blood is a relatively early phenomenon prior to death, so screening for NRBCs may aid in the early identification of patients at high risk. Further studies are needed to clarify whether the detection of NRBCs could help to decide on a change of patient management, but our present data suggest that NRBC-positive patients should obtain ongoing intensive care treatment.
The detection of NRBCs in the blood of medical intensive care patients is associated with significantly increased in-hospital mortality (50.7% versus 9.8%). The prognostic significance of NRBCs was independent of other laboratory and clinical risk parameters. The predictive value for death increased with the NRBC concentration and seems to decline again when the NRBCs have disappeared from the circulation.
Our present data suggest that NRBC-positive patients should obtain ongoing intensive care treatment.
- APACHE II:
Acute Physiology and Chronic Health Evaluation II
nucleated red blood cell
- SAPS II:
Simplified Acute Physiology Score II.
- Budmiger H, Graf C, Streuli RA: [The leukoerythroblastic blood picture. Incidence and clinical significance]. Schweiz Rundsch Med Prax 1984, 73: 1489-1493.PubMedGoogle Scholar
- Mettier SR: Hematologic aspects of space consuming lesions of the bone marrow (myelophtisic anemia). Ann Intern Med 1940, 14: 436-448.View ArticleGoogle Scholar
- Schwartz SO, Stansbury F: Significance of nucleated red blood cells in peripheral blood; analysis of 1,496 cases. J Am Med Assoc 1954, 154: 1339-1340.View ArticlePubMedGoogle Scholar
- Delsol G, Guiu-Godfrin B, Guiu M, Pris J, Corberand J, Fabre J: Leukoerythroblastosis and cancer frequency, prognosis, and physiopathologic significance. Cancer 1979, 44: 1009-1013. PublisherFullText10.1002/1097-0142(197909)44:3<1009::AID-CNCR2820440331>3.0.CO;2-JView ArticlePubMedGoogle Scholar
- Weick JK, Hagedorn AB, Linman JW: Leukoerythroblastosis. Mayo Clin Proc 1974, 49: 110-113.PubMedGoogle Scholar
- Tavassoli M: Erythroblastemia. West J Med 1975, 122: 194-198.PubMed CentralPubMedGoogle Scholar
- Andres WA: Normoblastemia after thermal injury. Am J Surg 1976, 131: 725-726. 10.1016/0002-9610(76)90188-4View ArticleGoogle Scholar
- Böning A, Stachon A, Weisser H, Krismann M, Skipka G, Laczkovics A, Krieg M: Postoperative cholesterol and erythroblasts as a parameter of perioperative mortality after cardiothoracic surgery. Z Herz-Thorax-Gefäâchir 2001, 15: 242-248. 10.1007/s003980170002View ArticleGoogle Scholar
- Otsubo H, Kaito K, Asai O, Usui N, Kobayashi M, Hoshi Y: Persistent nucleated red blood cells in peripheral blood is a poor prognostic factor in patients undergoing stem cell transplantation. Clin Lab Haematol 2005, 27: 242-246. 10.1111/j.1365-2257.2005.00687.xView ArticlePubMedGoogle Scholar
- Stachon A, Böning A, Krismann M, Weisser H, Laczkovics A, Skipka G, Krieg M: Prognostic significance of erythroblasts in blood after cardiothoracic surgery. Clin Chem Lab Med 2001, 39: 239-243. 10.1515/CCLM.2001.038View ArticlePubMedGoogle Scholar
- Briggs C, Harrison P, Grant D, Staves J, Machin SJ: New quantitative parameters on a recently introduced automated blood cell counter – the XE 2100. Clin Lab Haematol 2000, 22: 345-350. 10.1046/j.1365-2257.2000.00330.xView ArticlePubMedGoogle Scholar
- Walters J, Garrity P: Performance evaluation of the Sysmex XE-2100 hematology analyser. Lab Hematol 2000, 6: 83-92.Google Scholar
- Stachon A, Sondermann N, Krieg M: Incidence of nucleated red blood cells in the blood of hospitalised patients. Infusion Ther Transfus Med 2001, 28: 263-266. 10.1159/000050251Google Scholar
- Stachon A, Sondermann N, Imöhl M, Krieg M: Nucleated red blood cells indicate high risk for in-hospital mortality. J Lab Clin Med 2002, 140: 407-412. 10.1067/mlc.2002.129337View ArticlePubMedGoogle Scholar
- Stachon A, Holland-Letz T, Kempf R, Becker A, Friese J, Krieg M: Poor prognosis indicated by nucleated red blood cells in peripheral blood is not associated with organ failure of the liver or kidney. Clin Chem Lab Med 2006, 44: 955-961. 10.1515/CCLM.2006.183View ArticlePubMedGoogle Scholar
- Stachon A, Eisenblätter K, Köller M, Holland-Letz T, Krieg M: Cytokines and erythropoietin in the blood of patients with erythroblastemia. Acta Haematol 2003, 110: 204-206. 10.1159/000074228View ArticlePubMedGoogle Scholar
- Stachon A, Bolulu O, Holland-Letz T, Krieg M: Association between nucleated red blood cells in blood and the levels of erythropoietin, interleukin-3, interleukin-6, and interleukin-12p70. Shock 2005, 24: 34-39. 10.1097/01.shk.0000164693.11649.91View ArticlePubMedGoogle Scholar
- Stachon A, Kempf R, Holland-Letz T, Friese J, Becker A, Krieg M: Daily monitoring of nucleated red blood cells in the blood of surgical intensive care patients. Clin Chim Acta 2006, 366: 329-335. 10.1016/j.cca.2005.11.022View ArticlePubMedGoogle Scholar
- Knaus WA, Draper EA, Wagner DP, Zimmerman JE: APACHE II: a severity of disease classification system. Crit Care Med 1985, 13: 818-829. 10.1097/00003246-198510000-00009View ArticlePubMedGoogle Scholar
- Le Gall JR, Lemeshow S, Saulnier F: A new simplified acute physiology score (SAPS II) based on a European/North American multicenter study. JAMA 1993, 270: 2957-2963. 10.1001/jama.270.24.2957View ArticlePubMedGoogle Scholar
- Lehnhardt M, Katzy Y, Langer S, Druecke D, Homann HH, Steinstraesser L, Steinau HU, Krieg M: Prognostic significance of erythroblasts in burns. Plast Reconstr Surg 2005, 115: 120-127.PubMedGoogle Scholar
- West CD, Ley AB, Pearson OH: Myelophthisic anemia in cancer of the breast. Am J Med 1955, 18: 923-931. 10.1016/0002-9343(55)90173-XView ArticlePubMedGoogle Scholar
- Ruzicka K, Veitl M, Thalhammer-Scherrer R, Schwarzinger I: The new hematology analyzer Sysmex XE-2100: performance evaluation of a novel white blood cell differential technology. Arch Pathol Lab Med 2001, 125: 391-396.PubMedGoogle Scholar
- Stachon A, Holland-Letz T, Krieg M: High in-hospital mortality of intensive care patients with nucleated red blood cells in blood. Clin Chem Lab Med 2004, 42: 933-938. 10.1515/CCLM.2004.151View ArticlePubMedGoogle Scholar
- Stachon A, Lehnhardt M, Katzy Y, Holland-Letz T, Steinau HU, Krieg M: Making the case for adapting the abbreviated burn severity index to include erythroblast count. J Wound Care 2005, 14: 97-100.View ArticlePubMedGoogle Scholar
- Rosenberg AL: Recent innovations in intensive care unit risk-prediction models. Curr Opin Crit Care 2002, 8: 321-330. 10.1097/00075198-200208000-00009View ArticlePubMedGoogle Scholar
- Morgenthaler NG, Struck J, Christ-Crain M, Bergmann A, Muller B: Pro-atrial natriuretic peptide is a prognostic marker in sepsis, similar to the APACHE II score: an observational study. Crit Care 2005, 9: R37-R45. 10.1186/cc3015PubMed CentralView ArticlePubMedGoogle Scholar
- Schetz MR, Van den Berghe G: Do we have reliable biochemical markers to predict the outcome of critical illness? Int J Artif Organs 2005, 28: 1197-1210.PubMedGoogle Scholar
- Ratanarat R, Thanakittiwirun M, Vilaichone W, Thongyoo S, Permpikul C: Prediction of mortality by using the standard score systems in a medical intensive care unit in thailand. J Med Assoc Thai 2005, 88: 949-955.PubMedGoogle Scholar
- Herbert PC, Wells G, Blajchman MA, Marshall J, Marin C, Pagliarello G, Tweeddale M, Schweitzer I, Yetisir E: A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med 1999, 340: 409-417. 10.1056/NEJM199902113400601View ArticleGoogle Scholar
- Ward HP, Holman J: The association of nucleated red cells in the peripheral smear with hypoxemia. Ann Intern Med 1967, 67: 1190-1194.View ArticlePubMedGoogle Scholar
- Burkett LL, Cox ML, Fields ML: Leukoerythroblastosis in the adult. Am J Clin Pathol 1965, 44: 494-498.PubMedGoogle Scholar
- Clifford GO: The clinical significance of leukoerythroblastic anemia. Med Clin North Am 1966, 50: 779-790.PubMedGoogle Scholar
- Vaughan JM, Oxon DM: Leuco-erythroblastic anaemia. J Pathol 1936, 42: 541-563. 10.1002/path.1700420302View ArticleGoogle Scholar
- Okpara RA: The prognostic significance of leukoerythroblastic anaemia in Nigerians. East Afr Med J 1985, 62: 185-188.PubMedGoogle Scholar
- Retief FP: Leuco-erythroblastosis in the adult. Lancet 1964, 1: 639-642. 10.1016/S0140-6736(64)91456-4View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.