Prognostic value of hemogram derived indices in polytrauma patients

1 Introduction

Trauma is a major public health concern as it is the fourth leading cause of mortality worldwide and the first cause among individuals <40 years old (1). Each year, more than five million deaths result from severe trauma. These figures are comparable to those caused by smoking and exceed those caused by HIV (2, 3). Furthermore, the economic costs associated with trauma are considerable, accounting for $3,536 per trauma patient (4) and 1%−2% of the gross national product (GNP) in both low- and high-income countries (5). Globally the cost of road trauma is estimated to be approximately US $518 billion per year (6).

Despite significant progress in trauma management, with the establishment of specialized trauma centers and the development of practice guidelines, the avoidable mortality rate remains high, exceeding 20% (7, 8). Suboptimal initial management, often due to an underestimation of trauma severity, remains the main cause. Early identification of patients at increased risk of morbidity and mortality and their referral to specialized centers may significantly improve their prognosis.

In this context, several tools have been developed to assess trauma severity, some are specific to trauma, while others are more general. The injury severity score (ISS), derived from the abbreviated injury scale (AIS), is one of the most widely used tools for evaluating trauma severity. However, it has some limitations, particularly in cases of incomplete injury lesional assessment, and focuses only on the three most seriously affected areas of the body (9).

Consequently, the new injury severity score (NISS) was developed to better estimate overall trauma severity by considering the three most severe injuries, regardless of the anatomical regions affected. However, its performance is limited in complex traumas and severe traumatic brain injuries (9).

Similar findings have also been reported for general severity scores such as the Simplified Acute Physiology Score II (SAPS II) and the Acute Physiology and Chronic Health Evaluation (APACHE II and III). While these scores have demonstrated reliability in medical conditions, they tend to underestimate severity in trauma cases, as their constituent parameters often show little alteration in the initial phase (10).

Recently, simple indices such as red cell distribution width (RDW), neutrophil-to- lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR) have demonstrated prognostic potential in several fields, including oncology, cardiology, and pulmonology. They may also be useful in traumatology, but currently available results remain controversial (11–13).

This study aims to evaluate the prognostic value of these hemogram-derived indices in polytrauma patients admitted to intensive care units.

2 Materials and methods

2.1 Study design

This is a retrospective, observational, and analytical study carried out in the Department of Anesthesia and Intensive Care at Sahloul Teaching Hospital, Sousse, over a four-month period from March 15th to July 15th, 2024.

2.2 Study objective

This study aims to compare hemogram-derived indices to pre-existing scoring systems in outcome prediction of polytrauma patients.

2.3 Study population

2.3.1 Inclusion criteria

This study included all polytrauma patients admitted to the Department of Anesthesia and Intensive Care at CHU Sahloul, Sousse, between January 1, 2022, and March 30, 2024, with an ICU exceeding 5 days.

2.3.2 Exclusion criteria

The exclusion criteria included:

• Patients with conditions that could interfere with leukocyte counts, such as myelodysplastic syndrome, leukemia, or human immunodeficiency virus (HIV).
• Patients who received a blood transfusion within 2 weeks before the trauma.
• Polytrauma patients with associated burn injuries.
• Missing or incomplete medical records.
• Patients transferred to other intensive care units.

2.4 Data collection

Data collection was based on a dataset of 110 variables categorized into six groups: sociodemographic characteristics; trauma mechanism and injury assessment; severity evaluation using ISS, SAPS II, and APACHE II scores; management modalities, including surgical intervention, mechanical ventilation, sedation, osmotherapy, and blood transfusion; hemogram-derived indices such as RDW, NLR, and PLR, assessed on the first, third, and fifth day of hospitalization; and patient outcomes, including length of stay, complications, and mortality.

2.5 Definition of variables

RDW is a parameter that reflects the degree of heterogeneity in erythrocyte volume (commonly known as anisocytosis). It is derived from the erythrocyte histogram and expressed as either a standard deviation or a coefficient of variation (14, 15).

NLR is a biological marker of systemic inflammation. It corresponds to the ratio of neutrophils to lymphocytes. It was first described by Zahorec (16) in 2001 in oncology patients presenting infectious, surgical, or inflammatory stress.

PLR is a thrombo-inflammatory biomarker. It corresponds to the ratio of platelets to lymphocytes. It was first described by Temiz et al. in 2014 (17) as an independent predictor of cardiovascular mortality in patients with STEMI (acute ST-segment elevation myocardial infarction).

2.6 Data analysis

Data analysis was performed using SPSS software version 20. Qualitative variables were expressed as frequencies and percentages, while quantitative variables were presented as means ± standard deviation or medians and extremes depending on the normality of their distribution.

The analytical study focused on three dependent variables: mortality, infectious complications, and thromboembolic complications. The prognostic performance of the indices was assessed in three steps:

• Comparison of means using the Student's *t*-test and setting up the p-value at 0.05.
• Discriminative analysis using receiver operating characteristic (ROC) curves, with calculation of the area under the curve (AUC) to measure the ability of indices to differentiate between high- and low-risk patients. An AUC >70% indicates good discriminative performance, essential for effective clinical use in predicting complications and mortality.
• Assessment of scores cut-off, sensibility, specificity, positive predictive value (PPV), and negative predictive value (NPV).

2.7 Ethical considerations

Ethical considerations including approval of the institutional ethics committee, patient anonymity, and the confidentiality of personal data were respected.

3 Results

Between January 2022 and March 2024, 384 polytrauma patients were included in this study, with a mean age of 44.9 years and a male predominance (77.6%). The most common comorbidities were diabetes (10.9%), hypertension (10.7%), and obesity (4.4%). Road traffic accidents accounted for 78.1% of trauma cases. The most affected body areas were the thorax (86.7%) and the brain (57.6%) injuries. Upon admission, 20.6% of patients had a Glasgow Coma Scale (GCS) score <9. The mean severity scores were 29.5 for ISS, 22.5 for SAPS II, and 9.45 for APACHE II. The mean time to management was 3 h.

Biological analyses revealed mean RDW values of 13.87 at admission, 14.17 on day 3, and 14.64 on day 5. The NLR was 10.42 at admission, 10.63 on day 3, and 9.24 on day 5, while the PLR increased from 162.89 at admission to 168.79 on day 3 and 200.77 on day 5. The mean ICU length of stay was 14.17 days. Infectious complications occurred in 53.6% of patients and were dominated by pneumonia (49.2%) and bacteremia (21.6%). Thromboembolic complications affected 10.1% of patients. The overall mortality rate was 16.9%.

In the analytical study, indices significantly associated with mortality included RDW (at all time points), NLR (on days 3 and 5), and all four severity scores (ISS, Vittel, SAPS II, and APACHE II). However, no significant association was found for PLR. Results for other independent variables are summarized in Table 1. Regarding the predictive performance of these indices for mortality, the best discrimination power was observed for IGS II (AUC = 0.819) and APACHE II (AUC = 0.817), followed by RDW (AUC = 0.745). The indices with the weakest discriminative power were PLR and NLR (AUC = 0.489 and 0.494, respectively) (Figures 1–3). The sensitivity, specificity, PPV, and NPV of these indices are summarized in Table 2.

Table 1. Comparison of the means of different indices based on mortality between the two groups.

Bold values represent statistically significant results.

Table 2. Performance of different prognostic tools in predicting mortality.

Bold values represent statistically significant results.

For the association between these indices and the risk of septic complications, statistically significant differences were observed for RDW at all time points, as well as for NLR on days 3 and 5, and for ISS, Vittel, SAPS II, and APACHE II scores. However, PLR showed no significant differences at any time point (Table 3). Indices with an AUC ≥ 0.7, indicating good discriminative power, were SAPS II and APACHE II, particularly on days 3 and 5 of hospitalization. However, none of the hemogram-derived indices achieved an AUC ≥ 0.7 for predicting septic risk (Table 4).

Table 3. Comparison of the means of different indices based on the occurrence of infectious complications.

Bold values represent statistically significant results.

Table 4. Evaluation of the performance of different prognostic tools in predicting infectious complications.

Bold values represent statistically significant results.

The prognostic indices showed variable results in predicting thromboembolic complications. Significant differences were observed for NLR on day 3, as well as for SAPS II and APACHE II scores (Table 5). However, all these indices demonstrated limited performance, with AUC values below 0.7, indicating a modest discriminative capacity for predicting thromboembolic complications (Table 6).

Table 5. Comparison of mean values of different indices based on thromboembolic complications between the two groups.

Bold values represent statistically significant results.

Table 6. Evaluation of the performance of different prognostic tools in predicting thromboembolic complications.

4 Discussion

Trauma is a major public health problem because of its huge socio economic cost. Current scoring systems such as ISS and NISS may underestimate injuries severity leading to inadequate patient referral and suboptimal management. Indices derived from the hemogram, such as RDW, NLR, and PLR, show promising prognostic value in many fields. Recently, they have been assessed in trauma patients with conflicting results. This study aims to compare hemogram derived indices to preexisting scoring systems in the prediction of trauma patients' outcome.

Trauma related mortality is known to be due to “primary impacts” such as severe organ injuries, hypoxia, hypovolemia, and traumatic brain injuries. This initial phase triggers a systemic inflammatory response (SIRS), activating the immune and neuroendocrine systems. Subsequently, “secondary impacts,” such as infections or surgical procedures, may increase the inflammatory response, leading to multi-organ failure. SIRS is followed by a compensatory anti-inflammatory response syndrome (CARS) that may influence post-traumatic immunosuppression and infection risk by its intensity and duration. The release of inflammatory mediators, such as cytokines and alarmins, plays a key role in this process (18). The initial inflammatory response is triggered by DAMPs released during cell destruction, activating pattern recognition receptors and leading to the maturation of antigen-presenting cells, which can result in immunosuppression. This phase is clinically manifested by signs of SIRS, such as fever, tachycardia, and vasoplegia, with an increased risk of nosocomial infections (19). Additionally, activation of the hypothalamic-pituitary-adrenal axis and elevated cortisol levels contribute to managing SIRS but may also increase the risk of infections (20).

Neutrophils, which account for 60%−70% of circulating leukocytes, play a central role in the innate immune response and defense against infections following trauma while also contributing to tissue damage induced by inflammation. After trauma, neutrophilia is characterized by an increase in immature neutrophils and phenotypic heterogeneity, with a CD16bright/CD62Ldim immunosuppressive subtype, which is linked to complications such as multi-organ failure and acute respiratory distress syndrome (21, 22). Regarding lymphocytes, trauma induces T-cell depletion, contributing to post-traumatic immunosuppression and increasing vulnerability to infections. Persistent lymphopenia is associated with increased mortality and longer hospital stays (23, 24). Red blood cells (RBC) exhibit increased volume heterogeneity (RDW), which is related to inflammation and oxidative stress following trauma (25). Finally, platelets play a crucial role in coagulation, and their dysfunction, particularly in cases of hypothermia, can exacerbate hemorrhage.

Thrombocytopenia and platelet dysfunction are associated with higher mortality and increased transfusion requirements (26–29).

In this study, the mean age of patients was 44.96 ± 20 years, which is similar to the results of other international studies, such as those by Habibpour et al. (30) and Chae et al. (31). Men represented 77.6% of cases, which corresponds to a general trend observed in the literature due to men's increased exposure to road accidents and violence. Traffic accidents were the most common mechanism (78.1% of cases), as previously reported by Majercik et al. (32) and Dilektasli et al. (33).

The mortality rate was 16.9 % in this study while it ranged from 2% to 30% in the literature. Several factors such as injury severity, age, comorbidities, the speed of intervention, differences in trauma types, injury classification methods, and available medical resources play a crucial role in the observed outcomes.

Regarding the prognostic value of RDW in polytrauma patients, several studies have found similar results to ours. A prospective study conducted in China by Lv et al. in 2021 (34) showed that high RDW values at admission were significantly associated with early and late mortality in elderly patients undergoing hip fracture surgery. Average RDW was significantly lower in survivors than in non-survivors (13.2% vs. 14.54%; p <0.001) (34).

Similarly, an Austrian study published in 2020 by Weihs et al. (13) found a significant correlation between high RDW at admission and long-term mortality in polytrauma patients, with an RDW threshold value of 13.75%. According to the study of Kong et al. (35) elevated RDW at days 1 and 2 were predictive of short-term mortality in severe trauma patients, with AUCs of 0.705 and 0.720, respectively, suggesting that RDW may have a prognostic value.

However, several studies have reported different results. A retrospective study conducted in the United States by Majercik et al. and enrolling 2,512 patients, found no significant link between RDW at admission and mortality. Cox proportional hazard models did not find a statistically significant correlation between high RDW values at admission and mortality (p = 0.50). However, peak RDW values (>16.68) were associated with a higher mortality risk [HR = 2.72 (1.14–6.51)] (32). Similarly, Habibpour et al. (30) did not observe a correlation between RDW at admission and mortality in polytrauma patients with severe hemorrhages (p = 0.97).

Regarding the prognostic value of the neutrophil/lymphocyte ratio (NLR), mean NLR at days 3 and 5 were significantly associated with mortality and infectious complications in our study. However, its prognostic performances assessed by AUC of the ROC curve were not sufficient. Published results are also controversial concerning these points. The study of Vunvulea et al. (36) enrolling 611 polytrauma patients showed that high NLR values were strong indicators of post-traumatic pneumothorax and prolonged ICU stay, with an AUC of 0.798. Another study conducted by Vunvulea et al. (11) found a significant association between high NLR at admission and trauma severity measured by the ISS, with a coefficient of 0.448 (p = 0.001). Soulaiman et al. (37) also showed that high NLR on the first admission day was associated with increased mortality in polytrauma patients and overall survival was significantly lower in patients with an NLR > 4 (p = 0.020). Furthermore, a study conducted by Dilektasli et al. (33) found that NLRs higher than 8.19 on days 2 and 5 were independently associated with increased hospital mortality (p = 0.019 and p = 0.001, respectively).

However, in the study of El-Menyar (38), enrolling 1,199 abdominal trauma patients, there was no significant correlation between high NLR levels at admission and mortality. Similar results were published by Favre et al. (39) in their study including 216 thoracic trauma patients mainly regarding NLR performances in predicting delayed ARDS (p = 0.095).

Our study did not reveal any correlation between the platelet/lymphocyte ratio (PLR), measured at admission as well as on days 3 and 5 and early hospital mortality, infectious complications, or thromboembolic complications. In the study of Djordjevic et al. (12) conducted in Serbia, there was no significant association between PLR and mortality in severely traumatized patients (ISS > 25) and PLR were comparable between survivors and non-survivors (246.08 vs. 198.60, p = 0.356). However, several other studies have reported conflicting results. A recent study published in 2024 showed that PLR at admission to the emergency department was a significant biomarker for predicting short-term mortality in trauma patients. The ROC analysis revealed an optimal threshold of 65.35, with a sensitivity of 62.1% and specificity of 69.4% for predicting mortality (AUC = 0.742) (37). Another study conducted in Qatar in 2022 and enrolling 1,199 patients assessed the interest of PLR and NLR at admission in predicting mortality and massive transfusion in abdominal trauma patients. Results showed that PLR was significantly correlated with severity indicators such as the injury severity score, TRISS score, and serum lactate. PLR at admission had a sensitivity of 81.3% and a specificity of 61.1% for predicting mortality (38).

In our study, Hemogram derived indices have been compared to four trauma severity scores, namely ISS, Vittel criteria, SAPS II, and APACHE II. Considering mortality, the most effective scores were SAPS II (AUC = 0.819) and APACHE II (AUC = 0.817), followed by the RDW (AUC ranging from 0.745 to 0.792). Infectious complications were well predicted by the APACHE II (AUC = 0.730) and SAPS II (AUC = 0.728). No prognostic tool succeeds in predicting thromboembolic complications.

According to a study published in 2010, APACHE II was superior to SAPS II in traumatic brain injuries (40). A study enrolling 30,000 ICU patients in Île-de-France showed that SAPS II provided better calibration and discrimination (AUC = 0.93) for predicting mortality (41). However, its ability to predict hospital mortality may decrease with the length of stay. Concerning trauma-specific scores, a Tunisian study showed that ISS and NISS were the most effective, with AUCs of 0.929 and 0.892, much higher than those of SAPS II, TRISS, and RTS (10). Another study published in 2002 confirmed the superiority of ISS over NISS, with AUCs of 0.876 and 0.871, respectively (42).

However that may be, severity scores, as emphasized by the French society of anesthesia and intensive care, must be used with discernment and require periodic re-evaluation to remain relevant. Indices derived from the hemogram showed globally encouraging results in trauma despite the methodological limitations of the original studies. Better results could be obtained in the near future, especially as these indices have excelled in other fields.

In oncology, a study on rectal cancer patients showed that high NLR was associated with poor postoperative survival (43). In the context of COVID-19, a high RDW at admission was correlated with an increased risk of mortality, making it an important prognostic indicator (44). In cardiology, research has shown that high NLR and PLR values predicted higher mortality in patients undergoing coronary interventions, suggesting their utility in risk stratification systems (45). In patients with chronic obstructive pulmonary disease (COPD), higher RDW values were associated with complications such as right ventricular dysfunction and pulmonary hypertension, highlighting their potential role in monitoring COPD exacerbations (46). Finally, in intensive care, a study published in 2024 analyzed 256,387 septic patients in the ICU and found that high RDW (≥16%) was significantly associated with an increased risk of 30-day mortality (HR: 1.887). This association persisted across different RDW thresholds (12% to 20%) and in patients with severe sepsis, particularly in shock (HR: 1.731), on mechanical ventilation (HR: 1.735), with bacteremia (HR: 2.380), or requiring hemodialysis (HR: 1.979). These results suggest that RDW could be a simple and useful prognostic marker for assessing the severity of sepsis in intensive care settings (47).

In light of these results, hematological indices may perform better in other conditions outside of trauma. While RDW is a robust prognostic marker in medical critical illness, where it reflects cumulative inflammatory and oxidative stress, its utility in acute trauma may be limited by the primarily mechanical nature of injury and the kinetics of the inflammatory response. In fact, although the systemic inflammatory response in trauma is initiated within minutes of injury, evidenced by rapid DAMP release and early IL-6 elevation, commonly used biomarkers such as RDW may not reflect this surge immediately. RDW typically rises after 24 hours due to stress erythropoiesis and cytokine-mediated suppression of erythropoiesis. This temporal lag may limit its prognostic value when measured at admission, particularly compared to medical conditions where inflammation is already established at the time of assessment. This suggests that RDW may be most valuable in conditions where biological dysregulation, rather than physical injury alone, drives outcomes. Its performance can also be influenced by the comorbidity burden, which is more prevalent in the medical context, and the use of blood transfusions, which is more common in the surgical context.

Finally, before any definitive conclusion, some limitations should be considered. The inclusion criterion of an ICU stay exceeding five days may introduce selection bias by excluding both early in-hospital deaths and patients with milder disease courses, potentially skewing the prognostic performance of the evaluated indices. This criterion was nonetheless justified by the study's aim to enable serial assessment of these indices, as their predictive value is likely influenced by the timing of sample collection. Although this design precludes conclusions about their utility in predicting early mortality, the study meaningfully contributes to the existing literature by elucidating their role in forecasting later outcomes among prolonged ICU patients. Others limitations should also be considered, mainly a missing data risk, the lack of external validation, the absence of dynamic modeling, and the potential effects of interventions on serial hemogram values.

5 Conclusion

Hematological indices may have prognostic value in trauma, especially for the RDW. Despite the added value being limited beyond established scores, the integration of all these tools in multiparametric prognostic models using machine learning algorithms may lead to better performances in predicting trauma mortality.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. We used artificial intelligence to select the journals in which we can submit the article.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/femer.2026.1675944/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Theodore Chan, University of California, San Diego, United States

Reviewed by: Angela Cornelius, Fort Worth Office of the Medical Director, United States

Tamorish Kole, University of South Wales, United Kingdom