Nucleated red blood cell and its specific fluctuations are risk factors of 28-day and 90-day all-cause mortality in ICU patients: an observational cohort study

Background Little was known about the first occurrence time of Nucleated red blood cell (NRBC)and the persistent duration of the positive results in ICU patients, nor the dynamic changes of NRBC count and quantified predictive value of NRBC for mortality in ICU patients. We hypothesized that the persistent presence of NRBC in ICU patients would be associated with all-cause mortality and death patients would have specific NRBC fluctuations compared to survival ones. Methods A total of 224 newly admitted ICU patients in our hospital were collected and were followed up to 90 days after admission to ICU. Biological information, clinical characteristics and laboratory indicators and dynamic changes of NRBC count was compared between survival group and death group. Factors plausibly interact with both NRBC and outcome were included in logistic regression analysis to explore the risk factors of 28-day and 90-day mortality,ROC curve were drawn to determine the predictive value of NRBC for 28-day and 90-day all-cause mortality for ICU patients. Results NRBC was positively correlated with NRBC-positive duration days, APACHE II score, SOFA score, CCI, CRP, PCT and RDW. The 28-day mortality and 90-day were 15.0%,31.3%, 41.2%, 56.7% and 32.7%,52.9%, 59.4%, 80.0% in patients with 0/μL, 1–100/μL, 101–200/ μL and more than 200/μL NRBC,respectively.NRBC in the death group had a rapid upward trend before death, while those in the survival group were stable at low levels.NRBC was a robust predictor of 28-day and 90-day all-cause mortality following multivariable adjustment. The adjusted odds of 28-day and 90-day all-cause mortality in patients with more than 200/μL NRBC were 5.087(95% CI, 1.960-13.202) and 4.922(95% CI, 1.369-17.703), relative to patients without NRBC. Area under the curve(AUC) of predicting 28-day all-cause mortality and 90-day all-cause mortality by NRBC were 0.685 and

Conclusion In ICU patients especially sepsis patients and non-respiratory failure patients, the presence of NRBC is a robust predictor of 28-day all-cause mortality and 90-day allcause mortality. Specific NRBC fluctuations were found in death patients compared to survival ones.

Background
The essence of the Nucleated red blood cell (NRBC) is the non-denucleated red blood cells(RBCs), i.e. Immature RBCs [1]. NRBC also has a certain capacity of carrying oxygen, but it is significantly lower than mature RBC [2]. In addition, because NRBC is larger than mature RBC, it is easy to rupture when passing through narrow capillaries [3]. Under normal condition, NRBC can only exist in a small amount in the peripheral blood of newborn within one week of birth, but not in adult peripheral blood [4]. But in some pathological conditions, such as excessive erythropoiesis or the abnormal function of spleen,NRBC can not be removed from the blood circulation and appear in the peripheral blood. The common diseases or pathological processes leading to the increase of erythropoiesis mainly include massive hemorrhage, the increase of destruction of erythrocytes caused by various reasons, anemia, systemic inflammation or the existence of hypoxia, etc [5]. High portal vein pressure can lead to splenomegaly or damage of spleen immune function caused by autoimmune antibody [6,7]. Therefore, once the NRBC appears in the peripheral blood, it indicates that there is a disorder of physiological balance in the body.
It was reported that NRBC was related to the prognosis of various diseases, including pregnancy complications, neonatal-related diseases, severe infectious diseases, hematopathy, malignant tumors and immunological diseases [8][9][10].Many scholars have found that NRBC was related to the mortality of intensive care unit (ICU)patients, Stachon A,et al. [11]found that the NRBC positive rate in ICU patients was 20.0%, while the mortality rate of NRBC-positive patients was 21.1%, which was significantly higher than that of NRBC-negative ones, and the mortality rate increased with the increase of NRBC content. Menk M's study [12] on the patients with severe acute respiratory distress syndrome found that the positive rate of NRBC was 75.5%, while the mortality rate of NRBC-positive patients was 50.8%, and the NRBC content at the time of admission to ICU was an independent risk factor for mortality. A large sample studies showed that the 30day mortality rate of NRBC-positive ICU patients was 9.0%, and the 90-day mortality rate was 14.0%, and mortality increased with the increase of NRBC,90-day mortality rate of the patients with more than 300 / µ L NRBC is over 20.0% [13].
However, most of the above studies were limited to comparing the mortality of NRBCpositive and NRBC-negative groups or different NRBC content groups, but did not discuss the time of the first occurrence of NRBC and the persistent presence days of NRBC,nor the dynamic changes of NRBC count and quantified predictive value of NRBC for mortality in ICU patients. We sought to determine whether the persistent presence of NRBC in ICU patients would be associated with 28-day and 90-day all-cause mortality.We hypothesized that death patients would have specific NRBC fluctuations compared to survival ones.

Source population
We abstracted 358 cases of newly admitted ICU patients from August 2018 to April 2019 by the electronic medical records of taizhou hospital in zhejiang province, China.A total of 109 patients did not have NRBCs measured,12 patients did not have blood pressure measured , 5 patients died within 24 hours,6 patients did not take blood test after admission to the ICU were excluded.Thus,224 patients finally constituted the cohort.

Data sources
Biological and clinical datas on all patients were collected retrospectively from Laboratory information system(LIS)and Hospital information system(HIS).

Specimens and laboratory test
About 10 mL fasting venous blood samples of ICU patients within 24 hours after admitted were collected. 2 ml of them were placed in the vacuum tube containing EDTA-K2 anticoagulant (BD company), and automatically detected in the assembly line of Mindray BC-6800plus automated Hematology System.The remaining 8 mL were placed in two tubes with coagulant (BD company), and centrifuged at 3500rpm for 5 minutes after keeping at room temperature for 30 minutes for serum. The serum samples were tested for liver and kidney function, electrolyte, C-reactive protein (CRP), serum amyloid A (SAA) and procalcitonin (PCT) by Abbott C16000 automatic biochemical analyzer and Robas E601 automatic electrochemiluminescence analyzer,respectively.At the same time, 2ml arterial blood was drawn and placed in a special tube for blood gas analysis and detection by Robas b121 automatic blood gas analyzer.All samples should be mixed immediately after extraction.

Exposure of interest and score systems
The exposure of interest was the highest absolute NRBC count occurring after ICU initiation. The absolute NRBC count was determined via fluorescent flow cytometry using the Mindray BC-6800plus Automated Hematology System. The absolute NRBC count was reported as the number of NRBCs per microliter of blood, and was stratified as 0/μL,1-100/ μL, 101-200/μL and more than 200/μL. We utilized the charlson comorbidity index(CCI) to assess the burden of chronic illness, which is well studied and validated. CCI was put forward by Charlson in 1987. It refers to the injury and abnormality of other organs or tissues except the basic diseases, and carries out integral evaluation on the patients' complications, such as congestive heart failure, myocardial infarction, cerebrovascular disease, dementia, peripheral vascular disease, chronic obstructive pulmonary disease, peptic ulcer disease, connective tissue disease, diabetes, moderate Severe chronic kidney disease, hemiplegia, leukemia, lymphoma, solid tumor, liver disease, etc [14]. APACHEⅡwas used for evaluation of the severity of ICU patients.APACHE II score is the sum of acute physiology score, age score and chronic health score. Acute physiology score includes body temperature, heart rate, respiration and blood pressure, arterial oxygen pressure and pH, serum potassium (K), sodium (Na), creatinine (Cr), white blood cell (WBC) and Hematocrit (HCT), record the worst value of each index within 24 hours after entering ICU [15].We used SOFA score to determine the number of organs with failure. Index of respiratory, blood, circulation, nerve and kidney are involved in the SOFA score. The worst value of daily evaluation was recorded [16].
Respiratory failure was defined as arterial partial pressure of oxygen (PaO2) < 60mmHg.
Sepsis was defined as the positive result of blood culture within 24 hours after admitting in ICU.

End points
The outcomes were 28-day and 90-day mortality.The end point of our research was September 1, 2019.All the cohort had vital status present at 28-day and 90-day after ICU initiation.

Research methods
Biological,clinical characteristics and laboratory indexes of ICU patients were compared between survival group and death group at 28-day and 90-day to explore the risk factors of 28-day and 90-day all-cause mortality.
Highest absolute NRBC count of ICU patients during hospitalization and days in ICU when NRBC initially appeared in peripheral blood were recorded.Correlation between absolute NRBC count and NRBC positive duration were analyzed.And correlations between absolute NRBC count and other laboratory indexes were discussed to speculate the possible causes of NRBC in ICU patients.
28-day and 90-day all-cause mortality compared among different NRBC levels to explore the relationship between absolute NRBC count and all-cause mortality of ICU patients.
Dynamic changes of absolute NRBC count of ICU patients in survival group and death group were compared in order to explore the similarities and differences of NRBC fluctuations in the two groups.
Multivariate logistic regression was used to explore the risk factors of 28-day and 90-day all-cause mortality and ROC curve was utilized to determine the predictive value of NRBC for 28-day and 90-day all-cause mortality .

Statistical analysis
SPSS19.0 statistical software was used for data processing and analysis, and Graphpad prism 8 was used for mapping. Datas with the normal distribution were represented by mean ± SD, comparison between two groups was represented by t-test, and the comparison among multiple groups was represented by one-way ANOVA.Non-normal distribution datas were represented by Median (P25-P75), comparison between two groups was represented by Mann -Whitney U test, and comparison among multiple groups was represented by Kruskal-Wallis H test.Categorical covariates were described by n (%), comparison between two groups or among multiple groups using contingency tables and chi-square testing. Spearman correlation analysis was used to analyze the correlation between non-normal distribution datas.Multivariate logistic regression was used to screen the risk factors of mortality, and ROC curve was used to analyze the predictive value of NRBC for 28-day and 90-day all-cause mortality in ICU patients. P <0.05 was considered with statistically differences.

Biological and clinical characteristics of ICU patients
A total of 224 patients, 145 males and 79 females, aged 74 (60-82) years, were included.
According to the 28-day and 90-day survival status, patients were divided into survival group and death group.Statistical differences were found in proportion of chronic kidney disease, APACHⅡscore, SOFA score, PLT count and absolute NRBC count between 28-day death group and 28-day survival group,indexes in 28-day death group were higher than 28-day survival group except for PLT count(P < 0.05).
Proportion of acute respiratory failure,chronic kidney disease and malignant tumor, values of CCI, APACHEⅡscore, SOFA score, PCT and absolute NRBC count of 90-day death group were higher than 90-day survival group (P < 0.05)( Table 1) Table 4. Table 4 Correlation  Table 5. vs. 32.7%, P < 0.05). In patients with respiratory failure, sepsis, transfer from other departments and pneumonia, 90-day all-cause mortality of NRBC positive group was significantly higher than that of NRBC-negative group (P < 0.05),as shown in Fig. 2.

Relationship between NRBC counts and all-cause mortality in ICU
The 28-day and 90-day all-cause mortality rate of all patients were 27.2% (61/224) and 47.8% (107/224). There were significant differences in 28-day and 90-day all-cause mortality rate among the groups of 0 / µ L, 1-100 / µ L, 101-200 / µ L and > 200 / µ L NRBC(P < 0.05). The higher the NRBC count, the higher the mortality was (Fig. 3).   Most patients in ICU were with severe diseases and poor prognosis, and the probability of death in a short time was relatively high. At present, 28-day mortality or 90-day mortality are mostly used as the end point in the study of prognosis of ICU patients at home and abroad [19]. Lilly,et al [20]reported that the average length of stay in ICU was about 12.6 days, and chant C,et al [21] found that was 49 days, which were not the same. In this study, the average length of stay was 20.5 days, and 28-day mortality and 90-day mortality were set to explore the risk factors of ICU patients. Although both of them were short-term mortality, the 28-day mortality rate was closely related to patients' treatment, various medical operations and hospital infections [22,23], while the 90-day mortality rate was not only related to the influencing factors of 28-day mortality, but also related to patients' age, family care after discharge, nutrition and natural progress of the disease itself [19].
A meta-analysis in 2008 showed that the total in-hospital mortality rate of ICU patients was 11-45%, among which 6.3-37% died in ICU [24]. Rui,et al. believed that the mortality rate of ICU patients after discharge was between 6.4% and 40%, depending on the severity of patients [25]. Braber,et al. found that the in-hospital mortality rate of ICU patients was 16.3%, and the two-year mortality rate was 26.6% [26]. In this study, 224 inpatients in ICU were included in the retrospective study, of which 61 died at 28-day after admission, with an all-cause mortality rate of 27.2%.107 died at 90-day after admission, with an all-cause mortality rate of 47.8%, similar to the above studies. In addition, because our hospital was a tertiary general hospital with large flow of patients, ICU patients had certain representativeness.
We found that risk factors of 28-day and 90-day mortality of ICU patients were different.
CCI, the proportion of acute respiratory failure and malignant tumor were the risk factors of 90-day mortality, but they did not affect the 28-day mortality, which may be related to the natural progress and outcome of various complications, including tumors. However, patients with acute respiratory failure usually use ventilator to assist in maintaining respiration after tracheotomy in hospital. However, due to the limitation of conditions after discharge, most families were unable to assist in breathing, which affects the respiratory circulation of patients and increases the risk of hypoxia [27]. However, the proportion of chronic kidney disease, APACHE Ⅱ score, SOFA score and NRBC count were not only the risk factors of 28-day mortality, but also the risk factors of 90-day mortality.
Kidney is one of the important excretory organs of human body,and creatinine and blood potassium are excreted through kidney. Most patients with severe chronic kidney disease need hemodialysis instead of renal function to discharge waste [28]. Because of the high cost of hemodialysis, some patients are unable to bear, and there are complications such as infection and thrombus[29], so it has certain impact on the 28-day and 90-day mortality of ICU patients.APACHE Ⅱ score and SOFA score cover respiratory, circulatory, urinary, digestive system, etc., which are generally recognized as the scoring standards for evaluating the severity of patients' condition [30]. There were many studies proved that APACHE Ⅱ score and SOFA score were related to the mortality of ICU patients [31][32][33], which were consistent with this study.
Some scholars thought that NRBC was closely related to severe infection and was associated with prognosis, such as Desai S, who found that the mortality rate increased with the elevated of NRBC count, and NRBC might be a biomarker for the prognosis of surgical septicemia patients [37]; Minior VK,et al. found the fetal NRBC count increased after rats exposed to hypoxia for more than 24 hours, with low weight and the slow organ growth, suggesting that the production of NRBC was related to hypoxia and had a negative effect for fetal development [38]. However, there were few comparative studies on the dynamic changes of NRBC in survival and death patients. Therefore, this study aimed to explore the relationship between 28-day mortality and 90-day mortality of NRBC and ICU patients by collecting all kinds of clinical and laboratory indicators of ICU patients for comprehensive analysis.
In this study, we found that the majority of NRBC in the survival group was negative, and that in death group was > 200 / µ L. Zhang Shan,et al. found that for patients with APACHE II score > 21, there was a certain correlation between admission time and mortality [39].
The mortality of patients admitted during non-statutory working hours was relatively higher than that admitted during statutory working hours. Therefore, we speculated that the time of NRBC presenting positive results for the first time can also affect the mortality of ICU patients. It was found that more than 1 / 3 of the NRBC-positive ICU patients showed positive results on the second day of admission, and the 28-day all-cause mortality rate of the patients who were positive for the first time more than 10 days in ICU

Ethics approval and consent to participate
Approval for the study was granted by the Medical ethics committee of Taizhou Hospital of Zhejiang Province. Requirement for consent was waived because the data were analyzed anonymously.

Consent for publication
Not applicable.

Availability of data and materials
The dataset supporting the conclusions of this article is not available  Figure 1 Correlation between highest absolute NRBC count and NRBC continuous positive days *P value is obtained by Spearman correlation analysis. Relationship between NRBC content and all-cause mortality at 28 and 90 days in ICU Figure 4