High values of RDW are attributed to the increase in variability in red blood cell volume, which is caused by the release of some immature red blood cells from the bone marrow into the blood over the disease course. The mechanisms underlying the increase in RDW values in inflammation-related diseases remain unclear. Inflammatory factors can affect the maturation of red blood cells in the bone marrow through several ways, including by inhibiting the production and function of erythropoietin, affecting the metabolism of iron, and shortening the lifespan of red blood cells [22–24].
Some inflammatory factors, such as TNF-α, can cause iron deficiency and increased red blood cell phagocytosis in humans, and IL-6 and IL-1β can directly affect the lifespan of circulating red blood cells and inhibit the maturation of red blood cells. The action of these inflammatory factors can lead to the release of newly generated immature red blood cells from the bone marrow into the blood, resulting in elevated RDW values. Therefore, elevated RDW values can also reflect the severity of inflammation in the organism to some extent [12, 25–26]. The high stress state of the body in inflammatory states, poor nutritional levels, excessive activation of the renin–angiotensin system, and impairment of renal function affect the quality and lifespan of red blood cells to some extent, resulting in elevated RDW values [27–29].
In this study, we retrospectively evaluated pediatric patients with septic shock admitted to our PICU center in the past 7 years. To exclude factors that may have affected the RDW values in children admitted to the hospital, we excluded patients with hematologic diseases and a history of blood transfusion within 4 months prior to admission. We also excluded patients with septic shock caused by nosocomially acquired infections because of the possible impact of in-hospital pretreatment on RDW values and prognosis. Among the 153 patients included in this study, the median age was 10 months and there were only four cases of urinary tract infections. These findings confirm that septic shock induced by infection occurs mostly in infants and children and that shock caused by infections of respiratory and gastrointestinal origins is more common. Different from respiratory origins infection, gastrointestinal origins infection always characteristics with disease onset imperceptible, disease progress quickly and always need surgeons combined therapy. Some etiologies, such as severe diarrhea, intestinal obstruction, appendicitis perforated and intussusception could induce severe abdominal infection, and then quickly development of shock.
It was found that the age and weight of children in the group with high RDW values (> 15%) were significantly lower and that the morbidity and mortality rates were significantly higher than those in the other two groups. The hemoglobin and creatinine levels were also relatively low, which may be related to the low basal values in young patients.
There were some previous studies found that RDW was positively related with the growth of the age[30], but not significantly related to the gender differences, the increase of MCV with age was partly contributed to the increase of RDW[31]. But these studies only focus on adults, ignore the group of children. Unlike adults, the physiological status of children varies greatly at different ages, so we could not copy the data of adult. Perhaps, the basal RDW values in younger patients were relatively high, and the grouping may only screened out pediatric patients with septic shock who were younger and relatively more difficult to treat (Table 1). It is thus important to identify which factor has a large impact on prognosis and the causal relationship between age and high RDW values in a follow-up, large-scale cohort study.
To better analyze the risk factors for poor prognosis in children with septic shock, we collected the patients’ basic information and laboratory test results as well as some key data about the prehospital and post-admission treatments of the patients. For example, we collected data on antibiotic (β-lactam and carbapenem antibiotics) use in the 24-h period prior to admission; however, there was no significant difference between the survival and mortality groups. This finding may be related to the fact that the antibacterial spectrum of antibiotics used by the patients were mostly ineffective against pathogenic bacteria.
Among the 67 patients with positive blood cultures, 37 were gram-negative rods (e.g., 16 Escherichia coli and 15 Pseudomonas aeruginosa), and most were drug-resistant, making it difficult for common second-generation cephalosporins to effectively inhibit bactericidal activity. Some of the patients with shock were given CRRT during antishock treatments because of the presence of water and sodium retention or renal dysfunction. However, this treatment did not appear to have a significant impact on prognosis.
Most patients in the non-survivor group were hemodynamically unstable, we collected the highest index of vasoactive inotropic score (VIS) within 24h after the patient admitted into the hospital(VIS=[dopamine(µµg/kg/min)]+[dobutamine(µg/kg/min)]+[milrinone(µg/kg/min)×10]+[adrenaline(µg/kg/min)×100]+[anorepinephrine(µg/kg/min)×100])[32], the result indicated that the non-survivor group’s VIS index was much higher than that in the survival group. Statistical analysis showed that the VIS was highly correlated with patient mortality. Other objective indicators, such as blood lactate levels, coagulation, leukocyte count, and platelet levels, significantly differed between the mortality survival groups, and all trends were generally consistent with the results of previously published, large national and international studies [2, 33–36]. It should be noted that the duration of mechanical ventilation and length of hospital stay were significantly shorter in the mortality group than in the survival group. On the contrary, the mechanical ventilator-free hours in survivor group was significantly longer than non-survivor group. This finding is mainly attributed to the fact that most of the deaths due to septic shock occurred within 48–72 h after admission, and 46 of the 61 patients in the mortality group died within 72 h of admission. This does not imply that the patients in the mortality group did not require prolonged mechanical ventilation and hospitalization. Therefore, we did not include these factors in the multivariable logistic regression analysis.
The PELOD2 score is an extensively used international prognostic scoring system for critically ill children [37] that can be scored dynamically, and the results can be used to determine the morbidity and mortality rates based on a formula. Many studies [38–42] reported that its predicted outcome is more consistent with the actual morbidity and mortality rates. In this study, the PELOD2 score was found to have good application prospects in pediatric patients with septic shock, and the PELOD2 score within 24 h of admission was significantly higher in the deceased patients than in the patients who survived. Logistic regression analysis suggested that in addition to high RDW values (> 15%), the incidence of MODS and high PELOD2 scores were independent prognostic factors for poor prognosis in pediatric patients with septic shock.
Using the PELOD2 score at admission to predict the prognosis of pediatric patients with septic shock, it was found that the AUROC was as high as 0.786 (95% CI: 0.711–0.862), confirming its good predictive value. The PELOD2 scoring system mainly involves the five major systems of the body: the neurological, respiratory, circulatory, urinary, and hematological systems. The indicators incorporated in the hematological system are leukocyte counts and platelet levels. Although there was no difference between the AUROC for PELOD2 and PELOD2 + RDW, but RDW had good specificity(0.804) when the cutoff was 15.65, and it can be quickly get in clinic, so it is also valuable to predict prognosis especially when we diagnosed the patient was septic shock already(quickly diagnosis through patients’ medical history, clinic symptoms, clinical examination and general vital signs after admission).
This study had some limitations. First, this study was a single-center retrospective study, and although we collected clinical data of the past 7 years, the sample size was small. At the same time, some data we can not got, especially the distribution of RDW baseline value in different age groups in children, and the results may be somewhat biased; second, because it was a retrospective study, some confounding factors were not included, such as the levels of iron ions, erythropoietin, and various inflammatory factors at the time of patient admission, which, if included, would have a better illustrative value; third, because of the limitations one and two, we can not selected some novelty statistical methods, such as machine learning method that have already done well in signal pathway prediction, gene expression analysis, biomarkers selected and epidemiological investigation, etc[43–46].
Finally, we attempted to assess dynamic changes in RDW values after admission to determine the prognosis of pediatric patients with septic shock and to evaluate the effectiveness of antishock therapy. However, most of the patients were treated with blood product transfusions within 24–48 h of admission, and the exogenous blood product input would have a greater impact on human RDW levels. Thus, this analysis could not be conducted. Nevertheless, this study is one of the few attempts to evaluate the prognostic value of RDW in children with septic shock at home and abroad, and subsequent prospective, multicenter studies with large sample sizes will help confirm the value of RDW in pediatric septic shock.