1. Introduction

Managing critically ill patients with trauma in the intensive care unit (ICU) is difficult because of the complexity and severity of their injuries [1], and trauma-induced immunosuppression increases infections, complications, and mortality [2]. Physiologic severity scores, such as the Acute Physiology and Chronic Health Evaluation, Sequential Organ Failure Assessment, and Simplified Acute Physiology Score, predict outcomes and support risk stratification [3, 4, 5]. However, urgent settings may preclude timely collection of the comprehensive data they require, and the C-reactive protein-to-lymphocyte ratio (CLR) offers a practical alternative in such circumstances. Across conditions, elevated CLR is associated with worse short-term outcomes in severe acute respiratory syndrome coronavirus 2 infection—including higher risks of critical illness, mortality, and severe presentation in the emergency department [6]—and has also been shown to predict survival in non-small cell lung cancer and adverse outcomes in colorectal liver metastases [7, 8]. CLR also tracks acute pancreatitis severity and outcomes in dilated cardiomyopathy [9, 10], differentiates perforated versus acute appendicitis [11], and identifies postoperative surgical site infection after lumbar procedures [12] and periprosthetic joint infection [13]. As a simple, rapidly obtainable composite reflecting infection burden and host immunity, CLR can facilitate the early identification of high-risk patients and guide targeted therapy [10, 14]. In trauma, C-reactive protein (CRP) levels increase rapidly post-injury, reaching a peak approximately on day 3 and corresponding with the severity of trauma and tissue damage. Persistent elevation suggests complications such as infection [15]. Lymphocyte counts commonly fall in the acute phase, reflecting immune dysregulation. The trajectory of lymphocyte fluctuations within the first seven days is significant for predicting late prognosis in patients with severe trauma [16], and the shift in early lymphocyte count may be utilized to predict prognosis in patients with trauma [17]. By combining the inflammatory response indicated by CRP levels with immune status reflected by lymphocyte counts, the CLR may offer valuable prognostic information for patients with trauma.

Despite the extensive use of CLR as a prognostic marker in nontraumatic conditions, its role in trauma remains unexplored. In particular, there has been little investigation into the predictive significance of CLR in trauma patients with critical illness in the ICU. Only one study has identified CLR as a meaningful predictor of 30-day mortality in older individuals after hip fracture surgery [18]. This study confirmed CLR as a strong indicator of mortality, revealing that patients who succumbed within 30 days exhibited markedly elevated preoperative CLR compared with survivors [18]. In contrast, CLR showed limited discriminative power as a standalone predictor of mortality in individuals with traumatic brain injury, although a significant association with mortality was still observed [19]. This difference emphasizes the necessity for larger-scale research to determine the predictive value of CLR in ICU patients with general trauma. The purpose of this study was, therefore, to determine the predictive significance of CLR in critically ill trauma patients admitted to the ICU.

2. Materials and methods

3. Results

3.1 Enrollment of the study patient cohort

The study cohort consisted of 26,605 trauma patients from the Trauma Registry System, spanning the years 2016 to 2022 (Fig. 1). Of these, 24,193 were adults aged 20 years or older. There were 3540 adult patients that were hospitalized to the ICU. After removing patients with burns (n = 65), hanging injuries (n = 5), drowning (n = 1), missing lab data (n = 1480), and inadequate ISS data (n = 4), the final study group consisted of 1985 patients. Within this cohort, 217 deaths occurred, and 1768 patients survived.

Fig. 1.Enrollment process of adult trauma patients into the study cohort. ICU: intensive care unit; ISS: injury severity score.

3.4 Comparison of patients grouped by value of CLR

Fig. 2 shows the ROC curve analysis for CLR in predicting mortality. An optimal CLR cutoff value of 93.6 was determined. At this cutoff, the specificity was 0.874; however, the predictive performance of CLR was poor (AUC, 0.515; sensitivity, 0.230). As shown in Table 3, patients were stratified into high (≥93.6) and low (<93.6) CLR groups to compare demographics and clinical outcomes.

Fig. 2.Performance characteristics of the C-reactive protein/Lymphocyte ratio (CLR) for predicting mortality. AUC: area under curve.

Table 3.Comparative analysis of patients with high and low C-reactive protein/Lymphocyte ratio (CLR) based on the optimal cut-off value of 93.6.

		CLR	OR (95% CI)	p
		≥93.6 n = 272	<93.6 n = 1713
Male, n (%)	204 (75.0)	1092 (63.7)	1.71 (1.28–2.28)	<0.001
Age, yr (mean ± SD)	60.3 ± 19.4	56.7 ± 19.7	-	0.004
Comorbidities, n (%)
	CVA	20 (7.4)	77 (4.5)	1.69 (1.01–2.81)	0.042
	HTN	110 (40.4)	601 (35.1)	1.26 (0.97–1.63)	0.087
	CAD	20 (7.4)	159 (9.3)	0.78 (0.48–1.26)	0.302
	CHF	1 (0.4)	13 (0.8)	0.48 (0.06–3.70)	0.474
	DM	68 (25.0)	357 (20.8)	1.27 (0.94–1.71)	0.120
	ESRD	17 (6.2)	47 (2.7)	2.36 (1.34–4.18)	0.002
GCS, median (IQR)	15 (8–15)	15 (9–15)	-	0.106
AIS ≥2, n (%)
	Head/neck	194 (71.3)	1305 (76.2)	0.78 (0.59–1.03)	0.083
	Face	38 (14.0)	303 (17.7)	0.76 (0.53–1.09)	0.131
	Thorax	103 (37.9)	509 (29.7)	1.44 (1.11–1.88)	0.007
	Abdomen	57 (21.0)	324 (18.9)	1.14 (0.83–1.56)	0.427
	Extremity	130 (47.8)	622 (36.3)	1.61 (1.24–2.08)	<0.001
ISS, median (IQR)	20 (16–27)	20 (16–25)	-	0.048
Mortality, n (%)	50 (18.38)	167 (9.78)	2.09 (1.48–2.95)	<0.001
Mortality AOR*	-	-	1.71 (1.16–2.51)	0.007
Hospital stay (d)	24.4 ± 24.8	18.9 ± 16.1	-	<0.001
AIS: Abbreviated Injury Scale; AOR: adjusted odds ratio; CAD: coronary artery disease; CHF: congestive heart failure; CI: confidence interval; CRP: C-reactive protein; CLR: CRP (mg/L)/Lymphocyte (109/L); CVA: cerebral vascular accident; DM: diabetes mellitus; ESRD: end-stage renal disease; GCS: Glasgow Coma Scale; HTN: hypertension; IQR: interquartile range; ISS: injury severity score; OR: odds ratio; SD: standard deviation. *Mortality adjusted by age, ESRD, GCS, and ISS.

Patients with high CLR scores were more likely to be male (75.0% vs. 63.7%; p < 0.001) and older (mean age 60.3 vs. 56.7 years; p = 0.004). They also had a higher prevalence of comorbidities, such as CVAs (7.4% vs. 4.5%; p = 0.042) and ESRD (6.2% vs. 2.7%; p = 0.002). The high CLR group showed a greater incidence of significant thoracic injuries (AIS ≥2) (37.9% vs. 29.7%; p = 0.007) and extremity injuries (47.8% vs. 36.3%; p < 0.001). The ISS was higher in the high CLR group (median 20 vs. 20, p = 0.048) than in the low CLR group. Mortality was significantly higher in patients with high CLR scores than in those with low CLR scores (18.4% vs. 9.78%; p < 0.001), with an AOR of 1.71 (95% CI 1.16–2.51; p = 0.007). The patients with high CLR scores also had longer hospital stays (mean 24.4 vs. 18.9 days; p < 0.001) than those with low CLR scores. The Kaplan-Meier survival curves demonstrated a significant difference between the high and low CLR groups, with a Log Rank test p = 0.007 (Fig. 3).

Fig. 3.Kaplan-Meier survival curves of the high and low CLR groups.

4. Discussion

This study demonstrated that the CLR is a significant independent risk factor for mortality among patients with trauma admitted to the ICU. In these patients, a high CLR level correlates with longer hospital stays and increased mortality. However, CLR alone has poor predictive ability for mortality outcomes, as shown by its limited accuracy (AUC, 0.515; sensitivity, 0.230). Some studies in various clinical settings [9, 10, 12] have reported a strong predictive value of CLR; however, this result contradicts their findings. Trauma involves diverse acute injury mechanisms that trigger distinct inflammatory responses, unlike chronic conditions characterized by prolonged inflammation. Furthermore, complex and heterogeneous nature of trauma, often affecting multiple organ systems, makes it challenging for a single biomarker, such as CLR, to accurately predict outcomes across diverse cases. Various diseases have proposed different CLR cutoff values for diagnostic purposes. Tonduangu et al. [6] found that in patients with coronavirus disease 2019, CLR thresholds of 78.3 and 159.5 for predicting infection and mortality showed sensitivities of 79% and 48% and specificities of 47% and 70%, respectively. Unlike CLR cutoff values of 21.25, 30.835, and 0.45, which are respectively suggested for the Omicron BA.2.2 variant infection [20], acute pancreatitis [10], and acute appendicitis perforation [11], this study identified an optimal CLR cutoff of 93.6 for predicting mortality in ICU trauma patients. Considering that the predictive performance of CLR in patients with trauma in the ICU is poor (AUC, 0.515) and has low sensitivity (0.230) but high specificity (0.874), a high CLR may indicate specific inflammatory or infectious conditions and identify patients at high risk for mortality, while its utility as a standalone predictor of trauma mortality is limited. This study demonstrated that patients with high CLR require more focused care. By integrating CLR with other clinical variables, healthcare providers may be able to improve risk assessment and personalize interventions to improve patient outcomes.

Several factors may explain why CLR underperforms as a prognostic tool in trauma compared with its utility in other diseases. Trauma patients constitute a heterogeneous cohort exhibiting varied damage patterns and physiological responses, making a single inflammation-based index less universally predictive. After a major trauma, innate-like lymphocytes surge during the hyperacute response. Studies report increased natural killer (NK) and natural killer T (NKT) cells within hours, but the profound lymphopenia developing 4–12 h later and persisting beyond 48 h is linked to multiple organ dysfunction and mortality [19]. Those trauma patients with lymphocyte counts ≤0.5 × 10⁹/L at 48 h have an approximately 45% risk of mortality. In parallel, CRP, an Interleukin-6 (IL-6)-dependent acute-phase protein, rises rapidly after injury and peaks by day 3 [19]. This mirror image of CRP elevation and lymphocyte decline underlies the CRP-to-lymphocyte ratio, a composite biomarker reflecting inflammation and immune status. In addition, this study highlights the lack of significance of individual CRP and lymphocyte counts in the multivariable analysis. However, CLR combines these markers to capture the balance between systemic inflammation and immune competence. These data suggest that CLR may reflect the interplay of inflammatory burden and immune suppression better than either marker alone. Studies of specific diseases with more homogeneous patient groups have demonstrated that CLR is more effective compared with a diverse patient population [9, 10, 12]. This study discovered additional important characteristics linked to mortality, including age, GCS score, ISS, and preexisting illnesses, such as ESRD. These factors may have a more significant influence on patient outcomes than CLR alone [20]. Outcomes after trauma (especially severe injuries, such as traumatic brain injuries) depend heavily on injury severity and timely surgical or critical care interventions, which can overshadow systemic inflammatory markers [19]. In cases of isolated traumatic brain injuries, a high CLR correlated with mortality; however, its discriminative capacity was inadequate and became insignificant after correcting for injury severity [19]. This suggests that CLR in trauma largely reflects the underlying injury complexity rather than serving as an independent predictor. Furthermore, the acute immune responses to trauma are dynamic. Current evidence indicates that CLR in trauma primarily mirrors the severity of shock and tissue injury, rather than providing novel prognostic insights. In critically injured patients, a high CLR is often driven by an intense acute-phase reaction (marked by elevated CRP levels) coupled with stress-induced lymphocyte changes, which are hallmarks of severe hemorrhagic shock and trauma stress [21]. CRP levels rise within hours and peak around the third day post-injury [22], in proportion to tissue damage, whereas lymphocyte counts often decrease in the acute phase due to stress-induced immunosuppression. Such fluctuating kinetics imply that a single CLR measurement on admission may miss critical changes over time [19]. Indeed, persistent lymphopenia or secondary spikes in CRP levels can signal complications, such as organ failure or infection, later in the disease course [23, 24]. Reliance on the admission CLR value without accounting for its evolution or the timing of interventions is a potential limitation. In summary, the complexity of trauma, with variable injury profiles and rapidly evolving inflammation, likely blunts the prognostic precision of CLR.

This study has some limitations. First, this retrospective, single-center study may not generalize to all trauma populations or healthcare environments, and the potential for selection bias remains a concern. Second, the imbalanced number of survivors and deaths may have affected the findings and overall perception of the prognostic value of CLR. Third, laboratory data were dynamic and may have fluctuated during admission. In this study, laboratory data were derived from the initial analysis of blood samples collected upon admission to the emergency room. However, the time lag between arrival in the emergency room and admission to the ICU may vary, leading to bias in data assessment. Moreover, omitting patients with incomplete data, especially those lacking laboratory results, may have heightened bias if these individuals routinely differed from the included cohort. Treatment interventions (e.g., intubation or surgery) varied across patients with different injuries. In this study, we could infer that management differences did not significantly affect the outcomes. Finally, although the study concentrated on in-hospital mortality as the principal outcome, it neglected to investigate other significant clinical endpoints, like functional outcomes, quality of life, or long-term survival, which could offer a more comprehensive understanding of CLR levels in trauma patients. Addressing these limitations in future prospective multicenter research using consistent data collection techniques would assist in a deeper understanding of the CLR level and outcomes for trauma patients. Furthermore, the comparison of CLR with other ratios, such as neutrophil-to-lymphocyte ratio (NLR), PLR, and the CRP/albumin ratio, and the exploration of heterogeneity in subgroups (e.g., traumatic brain injuries, elderly, or poly-trauma patients) may provide additional information regarding the role of CLR in the trauma patients with critical illness.

5. Conclusions

This study demonstrated that a high CLR correlates with a longer hospital stay and increased mortality. However, CLR alone possesses little predictive value and should not be utilized as a solitary clinical indicator for traumatic patients with critical illness. The heterogeneity of trauma and dynamic immune responses likely explain why CLR alone underperforms. Future prospective, multicenter studies should incorporate serial CLR measurements and compare it with other inflammatory ratios to improve risk stratification.

Availability of data and materials

The raw data could be provided via corresponding author only for academic research.

Author contributions

PJK—writing. CYH—manuscript drafting. KHL—literature research. CHT—table creation. WTS—maintenance of the registered trauma database. SYH—statistical analysis. CHH—study design, proof-reading, and supervision.

Ethics approval and consent to participate

This study was approved by the Institutional Review Board (IRB) of the Chang Gung Memorial Hospital with approval number 202400890B0. The requirement for informed consent was waived by IRB because it entailed a retrospective review of registered trauma data.

Acknowledgment

We thank the Biostatistics Center at Kaohsiung Chang Gung Memorial Hospital for assisting us with statistical analyses.

Funding

This study was supported by a grant from Chang Gung Memorial Hospital CDRPG8M0022.

Conflict of interest

The authors declare no conflict of interest.