Association of early interleukin-6 level with injury severity and mortality in trauma patients: Systematic review and meta-analysis

Abstract

BACKGROUND

Polytrauma is associated with the release of various biomarkers, including interleukin (IL)-6, a proinflammatory cytokine with prognostic value.

AIM

To evaluate the association between IL-6 measured within the first 24 hours of admission and the injury severity score (ISS) among adult polytrauma patients and to explore the link between early IL-6 levels and in-hospital mortality.

METHODS

We conducted a systematic review and meta-analysis of articles published between 2004 and 2024 in the OVID EMBASE and PubMed databases. The primary outcomes of interest were ISS and all-cause in-hospital mortality. Heterogeneity among the included studies was assessed using the I² statistic. Correlations were performed using Pearson’s or Spearman’s correlation coefficients. The Newcastle-Ottawa Scale was used to assess bias in the included studies.

RESULTS

A total of 1126 studies were screened, of which 15 were selected, yielding 2106 polytrauma patients. The pooled effect size for the correlation between IL-6 and ISS using the Pearson’s coefficient was 0.49 [95%CI: 0.36-0.60; I²: 72.2%, P (heterogeneity) = 0.02], while that for studies using the Spearman’s coefficient was 0.50 [95%CI: 0.41-0.58; I²: 33.2%, P (heterogeneity) = 0.18]. Six studies indicated that IL-6 levels were significantly elevated in non-survivors compared with survivors.

CONCLUSION

Among polytrauma patients, elevated IL-6 levels within the first 24 hours of admission are associated with higher rates of severe injury and in-hospital mortality. Future research should investigate how early IL-6 levels can be used with other biomarkers to predict injury severity and worse outcomes. Additionally, larger-scale studies should be conducted to assess the tool's validity across a variety of populations.

Key Words: Serum interleukin-6; Polytrauma; Mortality; Injury severity score; Meta-analysis

Core Tip: Polytrauma is associated with increased morbidity and mortality. Use and identification of biomarkers as prognostic indicators in polytraumatized patients is therefore important. Interleukin (IL)-6 is a proinflammatory cytokine that is often elevated in polytrauma patients and is linked with worse outcomes and more severe injury. However, to date, there is only one meta-analysis published in English literature 7 years ago looking at the association between injury severity and IL-6. By integrating more recent evidence, this meta-analysis provides further support for IL-6 as a prognostic indicator not only for injury severity but also for mortality among adult polytrauma patients.

INTRODUCTION

Traumatic injury causes nearly 6 million deaths per year, surpassing the number of deaths caused by serious infectious diseases[1]. Polytrauma patients constitute almost 25% of hospitalized major trauma patients and tend to have higher risks of mortality and other adverse outcomes, highlighting the importance of identifying factors that predict mortality risk and severe injury[2].

Polytrauma is associated with the release of various biomarkers, including interleukin (IL)-6, a proinflammatory cytokine that has been found to be a prognostic indicator of the extent of trauma and worse outcomes[3]. This link with the severity of trauma may be related to the amount of tissue damage[4]. Studies have revealed that while IL-6 is initially released as part of the acute phase response to injury in order to promote healing of damaged tissues, polytraumatized patients tend to have prolonged IL-6 elevation, which may trigger severe inflammatory reactions and unregulated action on tissues far from the original site of injury, leading to morbidity and mortality[5,6].

Thus, understanding the role of IL-6 as an indicator of injury severity and mortality in polytrauma represents an important area of clinical research. Various studies have demonstrated associations between IL-6, injury severity, and mortality as shown in Table 1[7-21]. In a meta-analysis published in English 7 years ago[22], injury severity score (ISS) ≥ 9 or > 16 was significantly associated with elevated IL-6 levels. Therefore, we aimed to conduct a systematic review and meta-analysis to assess the association among IL-6 levels, injury severity, and mortality in adult polytrauma patients. ISS will be utilized as a continuous variable in this analysis.

Table 1 Summary and characteristics of the studies included in the systematic review and meta-analysis, mean ± SD/median (25^th-75^th percentiles).

Ref.	Study design	Country	Duration	Sample size total/gender (% male)	Age (years)	IL-6 within 24 hours (pg/mL) measuring time	ISS	Significant findings
Gołąbek-Dropiewska et al[7], 2018	Prospective cohort	Poland	NR	50/64%	39.5 (18-92)	10.65 (8.6-189.4) (24 hours)	26 (18-41)	Significantly higher levels of IL-6 on the first day after trauma in patients with ISS > 34
Binkowska et al[8], 2018	Prospective cohort	Poland	NR	32/71.9%	42.74 ± 18.41	ISS < 20: 42; ISS ≥ 20: 602 (at admission)	ISS > 20: n = 20; ISS < 20: n = 12	Statistically significant positive correlation between baseline levels of IL-6 and ISS (r = 0.64, P < 0.001)
Gupta et al[9], 2015	Prospective cohort	India	January 2010-January 2013	114/86%	34.3 ± 14.5	Survivors: 50 (10-90)2. Nonsurvivors: 250 (60-700)2, (24 hours)	18.71 ± 8.48	Serum levels of IL-6 on admission were significantly elevated in non-survivors (P < 0.05). IL-6 positively correlated with ISS (r = 0.6224, P < 0.0001)
El-Menyar et al[10], 2021	Prospective cohort	Qatar	October 2016-July 2019	250/98%	35.1 ± 10.1	137 (121-153)3 (at admission)	14.7 (13.4-15.9)3	Initial serum levels of IL-6 correlate with ISS (r = 0.40, P = 0.001)
Almahmoud et al[11], 2015	Retrospective case-control	United States of America	NR	472/69.9%	48.4 ± 0.9	ISS 1-15: 700; ISS 16-24: 1554; ISS > 24: 2176 (24 hours)	19.6 ± 0.5	Levels of IL-6 are elevated significantly in severe traumatic injury when compared to moderate and mild traumatic injury (P < 0.001)
Alper et al[12], 2016	Prospective cohort	Turkey	March 2014-July 2015	84/NR	ISS > 15: 42 ± 15.04; ISS ≤ 15: 38.75 ± 15.91	ISS > 15: 88.43 ± 8.10; ISS ≤ 15: 12.38 ± 7.52 (at admission)	NR	IL-6 levels are significantly elevated in patients with severe injuries (ISS > 15) compared to mild injuries (ISS ≤ 15) (P = 0.004)
Ebrahimpour et al[13], 2018	Prospective cohort	Iran	January 2013-December 2013	161/59.6%	39.28 ± 9.23	Died: 250.24 ± 21.96; survived: 251.27 ± 33.07 (24 hours)	29.24 ± 7.44	High levels of IL-6 on post-trauma day 1 are associated with high levels of ISS (P = 0.001). High levels of IL-6 on post-trauma day 2 are associated with mortality (P < 0.0001)
Akkose et al[14], 2007	Prospective cohort	Turkey	August 2003-May 2005	20/NR	36 ± 15	0.16 ± 0.06 (at admission)	24.8 ± 9.02	High IL-6 levels are associated with higher ISS (r = 0.448, P = 0.047). IL-6 is statistically insignificant in predicting mortality
Bogner et al[15], 2009	Retrospective	Germany	NR	58/69%	42 (18-89)	Patients receiving mass transfusion: 3752. Patients not receiving mass transfusion: 1502 (at admission)	35.4 ± 13.0	IL-6 was not significantly associated with higher ISS; however, patients who died within 90d after the trauma tended to have elevated IL-6 levels on admission and within the first 6 hours after trauma (P < 0.005)
Johansson et al[16], 2011	Prospective cohort	Denmark	2003-2005	75/7.7%	High syndecan-1: 45 (30-55)1. Low syndecan-1: 37 (29-48)1	High syndecan-1: 44.9 (13.5-95.5)1. Low syndecan-1: 7.8 (4.6-25.2)1 (at admission)	High syndecan-1: 23 (14-37)1. Low syndecan-1: 18 (14-28)1	IL-6 was correlated with ISS only in patients with high syndecan-1 (ρ = 0.41, P = 0.010)
Yagmur et al[17], 2005	Prospective cohort	Turkey	December 2003-April 2004	99/71.7%	Survivors: 25 ± 21. Non-survivors: 32 ± 22	ISS < 16: 134 ± 71. ISS ≥ 16: 202 ± 64 (24 hours)	Survivors: 9.8 ± 5.1. Non-survivors: 19.8 ± 12.7	Patients with ISS > 16 had higher IL-6 levels than patients with ISS ≤ 16 (P < 0.001). Patients who died from trauma also had elevated IL-6 (P = 0.05) compared to those who survived
Sousa et al[18], 2015	Prospective cohort	Portugal	January 2010 to December 2010	99/82.8%	31 (18-60)1	Admitted to ICU: 569 (177-1440). Not admitted to ICU: 190 (131-439)1 (24 hours)	29 (17-52)1	On-admission serum IL-6 levels correlated with ISS (r = 0.346, p not reported)
Taniguchi et al[19], 2016	Prospective cohort	Japan	March 2014 to December 2014	208/74%	ICU > 7 days: 58 (43-71)1. ICU ≤ 7 days: 40 (20-59)1	ISS 1-3: 302. ISS 4-8: 352. ISS 9-15: 552. ISS 16-24: 752. ISS ≥ 25: 1702 (at admission)	ICU > 7 days: 26 (17-35)1. ICU ≤ 7 days: 7 (1-14)1	On-admission IL-6 levels correlated with ISS (r = 0.459, P < 0.0001). Patients who died within 28 days had significantly higher IL-6 levels than survivors (P = 0.021)
Stensballe et al[20], 2009	Prospective cohort	Denmark	2004-2005	265/72.1%	38.1 (26-55)1	30 (10-79.3)1 (12 hours)	Survivors: 9 (4-16)1. Non-survivors: 38 (25-75)1	IL-6 correlated significantly with ISS (on-admission: r = 0.52, P < 0.0001; 6h: r = 0.69, P < 0.0001). Serum IL-6 was higher in patients who died within the first 30 days (P < 0.0001)
Laishram et al[21], 2024	Prospective cohort	India	November 2022-June 2024	119/85.7%	NR	Inapparent hypoxia: 61.84 (39.41-158.25)1. Sepsis: 125.72 (43.05-421.77)1. MODS: 270.87 (48.28-496.95)1. FES: 50.74 (IQR NR) (12 hours)	ISS < 9: n = 38. 9-15: n = 64. ISS 16-25: n = 9. ISS ≥ 25: n = 8	12 hours serum IL-6 levels correlated with ISS (r = 0.595, P < 0.001)

¹Reported as median (25^th-75^th percentiles).

²Data was reported in graph, precise values not available.

³Reported as median (95%CI).

MODS: Multiple organ dysfunction syndrome; FES: Fat embolism syndrome; IL: Interleukin; ISS: Injury severity score; IQR: Interquartile range; ELISA: Enzyme-linked immunosorbent assay; ICU: Intensive care unit; NR: Not reported.

MATERIALS AND METHODS

This systematic review was conducted according to the PRISMA Statement and registered at the International Prospective Register of Systematic Reviews (PROSPERO on October 27, 2024; CRD42024600890).

Data sources and searches

A systematic review search was conducted in the OVID EMBASE and PubMed databases. A list of the search strategy for both databases is available in the Supplementary material. We also reviewed the reference lists of included studies to identify additional papers. All OVID EMBASE and PubMed studies were transferred to the systematic review management system Rayyan (Cambridge, MA, United States), and duplicates were removed. We limited the search to studies published in English and included only adult participants (aged ≥ 18 years) to prevent translational errors and ensure comparable physiology across patients (Supplementary material).

Inclusion criteria: Adult (≥ 18 years of age), polytrauma, serum or plasma IL-6 levels measured within the first 24 hours of admission, articles published between September 2004 and September 2024, randomized controlled trials (RCTs), observational studies, and manuscripts in English.

Exclusion criteria: Pediatric populations (< 18 years of age), isolated trauma, missing data on ISS and IL-6 within the first 24 hours of admission, review articles, case studies, case series, and non-English manuscripts.

Participants: Adult polytrauma patients.

Intervention: Serum or plasma IL-6 levels measured within the first 24 hours of admission.

Control or comparison: None. Polytrauma was defined as an abbreviated injury scale ≥ 3 in at least two body regions. Any studies that focused solely on trauma associated with a single body region were excluded.

Outcomes: ISS and all-cause in-hospital mortality.

Data extraction: Both reviewers (Naduvilekandy M and Al-Hassani I) independently screened the articles by title and abstract, then screened the full texts against the inclusion and exclusion criteria. Naduvilekandy M and Al-Hassani I cross-checked their decisions at each stage and resolved disagreements. Any unresolved conflicts were resolved by a third party (El-Menyar A). Data extracted from the studies included authors, the location of the study (country), year of publication, study duration, sample size, mean/median age, gender, type of trauma (blunt or penetrating), injury location, timing of IL-6 sampling, mean/median IL-6 and ISS, correlation data between IL-6 and ISS (or IL-6 level comparison between ISS subgroups), and mortality.

Methodological quality

The Newcastle-Ottawa Scale (NOS) was used to assess bias in the included studies. Both reviewers (Naduvilekandy M and Al-Hassani I) independently evaluated the studies and cross-checked their decisions. A third party (El-Menyar A) resolved any disputes between the two.

Statistical analysis

Meta-analysis was performed to estimate the pooled effect of the correlation between serum/plasma IL-6 levels and the ISS. Separate analyses were conducted for studies reporting the Spearman and Pearson correlation coefficients.

Fisher’s z-transformation was applied to the correlation coefficients to normalize their distribution, and the corresponding variances were calculated. A random-effects model using the DerSimonian-laird method was used to account for heterogeneity across studies. The pooled estimates of Fisher’s z values were then back-transformed to correlation coefficients for interpretability. Heterogeneity among the included studies was assessed using the I² statistic, with thresholds for low (0%-25%), moderate (26%-50%), substantial (51%-75%), and high (> 75%) heterogeneity. A funnel plot was generated to assess potential publication bias.

Heterogeneity was explored using sensitivity and subgroup analyses based on trauma phenotype (blunt vs mixed trauma), study quality, and sample size. Meta-regression and additional subgroup analyses (including hemorrhagic shock cohorts, assay platform, and sampling time) were not performed when analyses were underpowered or data were insufficiently granular. Leave-one-out sensitivity analysis was conducted to assess the influence of individual studies. Statistical analyses and visualizations were conducted using the meta and metafor packages in RStudio (Version 2024.12.0).

RESULTS

A total of 1365 records were identified through database searching. After removing 239 duplicates, 1126 records remained, of which 1091 were excluded for being unrelated to the topic. This yielded 35 full-text articles for eligibility assessment. Upon detailed review, 20 articles were excluded due to reasons such as measuring IL-6 beyond 24 hours (n = 3), incorrect IL-6 measurement (n = 2), lack of ISS measurement (n = 1), absence of comparison or correlation data (n = 2), including non-polytrauma populations (n = 6), being available only as abstracts (n = 2), or falling outside the period of interest (n = 4). Ultimately, 15 studies were included in the qualitative synthesis. The selection process is summarized in the PRISMA flowchart (Figure 1).

Figure 1  Flow chart of the study selection process.

Characteristics of included studies

Two of the included studies were retrospective, while the remaining were prospective. Seven studies were conducted before 2015, and eight were conducted afterward. Geographically, three studies originated in Turkey, two in Denmark, Poland, and India, and one in the United States, Portugal, Japan, Iran, and Qatar. Study durations ranged from a few months to several years, reflecting variability in design and data collection periods. Short-term studies included those lasting four months (e.g., from December 2003 to April 2004)[17] or a single year (e.g., from January 2013 to December 2013)[13], while longer studies spanned multiple years, such as from August 2003 to May 2005[14] and from October 2016 to July 2019[10]. Some studies provided only broad time frames (e.g., 2003-2005)[16], while others did not specify the duration[7,8,11,15]. Detailed timelines are summarized in Table 1.

The sample sizes of the included studies varied widely, ranging from 20 to 472 participants. Some studies enrolled relatively few participants, such as 32, 50, or 58, while others enrolled larger cohorts, including 99, 114, 150, or 208 participants. Meanwhile, a few studies had moderate sample sizes, such as 75, 99, or 119 participants. Most studies lacked a control group, although some studies included comparisons of healthy volunteers admitted to the hospital (non-trauma)[7,12,14,17].

The primary outcomes in the studies included IL, such as IL-6, and other inflammatory markers, such as syndecan-1 and p38 mitogen-activated-protein kinase. Injury severity was another common primary outcome and was often analyzed in relation to these inflammatory markers. Secondary outcomes in the studies included injury severity, multiorgan dysfunction syndrome (MODS), complications, and mortality, which were all evaluated in relation to IL-6 levels. For example, IL-6 levels were measured in relation to injury severity and adverse clinical outcomes, including prolonged intensive care unit stay, mortality, and infection complications. Notably, few studies have examined markers such as mean platelet volume.

The timing of IL-6 measurements varied across the included studies. All studies measured IL-6 levels within the first 24 hours of admission, with measurements taken at specific time points, including admission, within the first 6 hours, and at 12 and 24 hours. A few studies measured IL-6 levels at multiple time points, including 0, 3, 6, 12, and 24 hours[8,21].

The methods used to measure IL-6 levels varied across studies. The most commonly used method was the enzyme-linked immunosorbent assay (ELISA)[8,10,12,14,16,18-20]. Other methods used the IMMULITE[13,15], the Biomedica Immunotech EIA[17], and the Cytometric Bead Array Kit[9]. One study employed the MILLIPLEX MAP Human Cytokine/Chemokine Panel[11]. Fluorescence flow cytometry[7] and electrochemiluminescence[21] were also used in some studies.

The lowest measured IL-6 level was 0.16 pg/mL[14], and the highest was 569 pg/mL[18]. The lowest ISS reported was 7[19], and the highest ISS reported was 38[20].

Correlation between IL-6 and ISS

The included studies have reported the correlation between IL-6 and ISS using various approaches. Some utilized Spearman’s rank correlation coefficient[8,14,16,18,20-21], while others adopted Pearson’s correlation coefficient[9,10,19]. Further, some studies compared IL-6 levels across patient groups categorized by ISS values as shown in Table 1[7,11-13,15,17].

Across the six studies reporting Spearman correlation coefficients, the correlation between IL-6 levels and ISS was variable but consistently positive. At admission, moderate correlations with coefficients of 0.448 (P = 0.047) and 0.346 (P = not significant) were reported in two studies[14,18]. In contrast, another study showed a similar correlation only in patients with elevated syndecan-1 (r = 0.41, P = 0.010)[16]. Laishram et al[21] found that IL-6 levels at 12 hours of admission showed a moderate-to-strong correlation with ISS (r = 0.595, P < 0.001), and that levels at 24 hours showed a stronger correlation (r = 0.725, P < 0.001). Spearman’s rho values ranged from 0.346 (P value not reported) to 0.725 (P < 0.001) across the studies. Three studies reported Pearson correlation coefficients ranging from r = 0.40 to 0.622, all with significant P < 0.001[9,10,19]. Five studies reported neither a Pearson nor a Spearman rank correlation coefficient but demonstrated that IL-6 levels tended to be higher in cohorts with higher ISS[7,11-13,17]. However, one study reported that IL-6 was not significantly associated with higher ISS[15].

The pooled effect size for the three studies reporting a Pearson correlation between serum/plasma IL-6 levels and ISS was 0.49 (95%CI: 0.36-0.60), indicating a moderate positive correlation. Similarly, for the six studies reporting Spearman correlations, the pooled effect size was 0.50 (95%CI: 0.41-0.58), indicating a moderate positive correlation between IL-6 levels and ISS.

Pearson correlations exhibited substantial heterogeneity, with an I² value of 72.2%, indicating considerable variation across studies (Figure 2A). Subgroup or sensitivity analysis was not possible due to the limited number of studies. In contrast, Spearman correlations showed moderate heterogeneity (I² = 33.2%), suggesting less variation across studies (Figure 2B). Funnel plots for the correlation analyses (Figure 3) were generated to assess publication bias.

Figure 2 Forest plot. A: Using the Pearson correlation coefficient to assess the correlation between interleukin (IL)-6 and injury severity score (ISS); B: Using the Spearman rank correlation coefficient to assess the correlation between IL-6 and ISS.

Figure 3 Funnel plot. A: For Pearson’s coefficient (Egger’s P-value = 0.007); B: For Spearman’s rank coefficient (Egger’s P-value = 0.91).

IL-6 as an indicator of mortality

Seven studies assessed the association between IL-6 levels and mortality. Six of them reported a statistically significant elevation in IL-6 levels in non-survivors compared with survivors at one or more time points[9,13,15,17,19,20]. Only Akkose et al[14] found no significant difference between the groups (Table 2).

Table 2 Interleukin-6 values of survivors and non-survivors, mean ± SD.

Ref.	Survivors		Non-survivors		Significant difference	Time of IL-6 measure
	No. patients	IL-6 (pg/mL)	No. patients	IL-6 (pg/mL)
Akkose et al[14], 2007	14	0.15 ± 0.05	6	0.18 ± 0.08	> 0.05	On arrival
Stensballe et al[20], 20091	236	28 ± 5	29	200 ± 65	< 0.05	On arrival
						12 hours
Taniguchi et al[19], 20161	201	60 ± 10	7	420 ± 165	< 0.05	On arrival
Yagmur et al[17], 2005	21	86 ± 77	17	146 ± 134	0.05	On arrival
Bogner et al[15], 2009	47	NR	11	NR	< 0.05	On arrival, 6 hours
Ebrahimpour et al[13], 2018	136	251.27 ± 33.07	25	250.24 ± 21.96	> 0.05	24 hours
	136	223.53 ± 25.33	25	276.84 ± 12.51	< 0.05	48 hours
Gupta et al[9], 20151	77	70 ± 82	37	346 ± 445	< 0.05	On arrival
	77	96 ± 75	26	653 ± 576	< 0.05	72 hours

¹Approximated values from the figure.

NR: Not reported; IL: Interleukin.

Stensballe et al[20] observed significant differences in IL-6 levels at arrival and at 6, 12, and 24 hours post-admission. Similarly, Gupta et al[9] reported significant differences on days 0, 3, 7, and 14. Meanwhile, Ebrahimpour et al[13], who measured IL-6 levels across multiple time points (days 1, 2, 3, 4, 7, 10, and 14), found a significant difference between survivors and non-survivors only on day 2.

The pooled mean difference in IL-6 levels between survivors and non-survivors was significant (MD = -163.48; 95%CI: -291.90 to -35.06), with I² = 98.4% (Figure 4). No subgroup or sensitivity analysis substantially reduced heterogeneity.

Figure 4  Forest plot of the mean difference of interleukin-6 between survivors and non-survivors on arrival.

Methodological quality

The methodological quality of the studies was assessed using the NOS (Table 3). Scores ranged from 5 to 9, with all included studies demonstrating moderate to high quality. For studies without a comparator group, the “Selection of the non-exposed cohort” domain was not applicable, and they were scored on a scale of 8 instead of 9. All studies scored 3 or 4 in the selection section, indicating a clear definition of trauma populations and a reliable evaluation of exposure. Similarly, most studies scored 2 or 3 in the outcome domain, as IL-6 levels were commonly measured using validated techniques such as ELISA. However, comparability scores were low in many studies, with several receiving a score of 0 because they lacked adjustment for confounding variables.

Table 3 The quality of studies included in the systematic review and meta-analysis was assessed according to the Newcastle-Ottawa Scale for cohort studies.

Ref.	Selection	Comparability	Outcome	Score
Gołąbek-Dropiewska et al[7], 2018	4	0	3	7/8
Binkowska et al[8], 2018	4	2	3	9/9
Gupta et al[9], 2015	3	0	3	6/8
Stensballe et al[20], 2009	3	2	2	7/8
El-Menyar et al[10], 2021	3	2	3	8/8
Almahmoud et al[11], 2015	3	2	2	7/8
Alper et al[12], 2016	4	0	2	6/9
Ebrahimpour et al[13], 2018	3	0	3	6/8
Akkose et al[14], 2007	4	0	2	6/8
Bogner et al[15], 2009	3	0	3	6/8
Johansson et al[16], 2011	3	2	2	7/9
Yagmur et al[17], 2005	4	1	2	7/9
Sousa et al[18], 2015	3	0	3	6/8
Taniguchi et al[19], 2016	3	0	2	5/8
Laishram et al[21], 2024	2	0	3	5/8

DISCUSSION

To our knowledge, there is no prior meta-analysis investigating the association between IL-6 levels and injury severity in polytrauma as a primary outcome. However, a systematic review published seven years ago reported the association between IL-6 and posttraumatic multi-organ failure and mortality[22]. Therefore, this analysis aimed to synthesize the results of more recent literature examining IL-6 as a prognostic indicator of injury severity and mortality in polytrauma patients. Generally, polytrauma patients reach peak blood IL-6 levels within the first 12 hours after injury[20]. This early rise of IL-6, alongside its role in post-traumatic severe inflammatory reactions, has made this cytokine a promising predictor for worse outcomes in multiple trauma patients.

In the present meta-analysis, we observed a moderate positive correlation between IL-6 levels and ISS. Even when the association is stratified by ISS value, IL-6 levels tended to be higher amongst the cohorts defined at higher ISS values (Table 1)[7,11-13,17]. For example, Almahmoud et al[11] reported that patients with mild injuries (ISS 1-15) had IL-6 levels of 700 pg/mL, those with moderate injuries (ISS 16-24) had levels of 1554 pg/mL, and those with severe injuries (ISS > 24) had levels of 2176 pg/mL. This stepwise increase reflects that early IL-6 measurements can be used to triage into groups based on their ISS. Bogner et al[15] was the only study that did not follow this trend. The authors reason that this may be because they had set an ISS cut-off too high at 35; however, another study with a cut-off of 34 found a significant association between IL-6 and ISS[7]. Interestingly, Johansson et al[16] found that IL-6 correlated with ISS only in patients with high syndecan-1 levels. After trauma, tissue injury, and hemorrhage lead to endothelial glycocalyx degradation and the release of syndecan-1[23]. Studies in mice showed that increasing syndecan-1 levels did not alter proinflammatory cytokine levels, such as IL-6[24]. Thus, further research is needed to determine whether syndecan-1 modifies this relationship or mediates the association between IL-6 and ISS. Nonetheless, our findings are corroborated by an earlier meta-analysis, which found that higher IL-6 levels were found in patients with ISS ≥ 9 and ISS ≥ 16 compared to their lower ISS counterparts[22]. As such, an overwhelming body of evidence suggests that IL-6 may be an effective tool for assessing the extent of injury in polytrauma patients.

The present meta-analysis also found a statistically significant elevation in IL-6 levels among both non-survivors and survivors. Akkose et al[14] was the only study in this meta-analysis to report no significant difference in IL-6 levels between survivors and non-survivors. However, their sample size was very small (n = 20). Mechanistically, traumatic injury induces the release of damage-associated molecular patterns (DAMPs), which trigger a severe inflammatory response syndrome (SIRS) by unregulated, systemic activation of the innate immune system[25]. DAMP-activated innate immune cells release various proinflammatory cytokines, including IL-6, leading to elevated IL-6 levels in the blood. This leads to post-injury hyperinflammation, tissue hypoperfusion, organ dysfunction, and hence an increased risk of mortality[6]. A meta-analysis by Qiao et al[22] found that concentrations of IL-6 were significantly higher amongst those who suffered MODS and amongst those who died compared to patients without MODS and survivors, respectively.

Furthermore, numerous studies have identified cut-off IL-6 values that can be used to predict morbidity and mortality in trauma patients[18,26,27]. For example, Rao et al[27] found that on-admission IL-6 had a sensitivity of 75% and a specificity of 84.6% for predicting MODS in polytrauma patients, with a cut-off of 106.56 pg. However, these studies[18,26,27] report low sensitivities when IL-6 is used to predict outcomes. Thus, while most of the evidence indicates that IL-6 may be associated with mortality and adverse outcomes in polytrauma patients, future research should focus on evaluating models that incorporate IL-6 alongside other biomarkers to improve sensitivity.

Despite these associations with worse outcomes and mortality, it has been argued that worse outcomes may be confounded by severe injury rather than by the pathophysiology of IL-6 and SIRS[22]. While a higher ISS is often associated with adverse outcomes, IL-6 is easily measured on admission. In comparison, ISS cannot be definitively attained on arrival as it requires detailed radiological and operative evaluation[28]. Moreover, access to ISS data may take at least 48 hours to be available in the patient’s electronic medical record; therefore, IL-6 is more valuable than ISS early post-admission. Additionally, Stensballe et al[20] reported that IL-6 levels are higher among non-survivors than among survivors from admission through the first 24 hours in the hospital. Stensballe et al[20] also found a similar relationship in injury severity, with higher ISS associated with higher IL-6 levels on arrival and after 6 hours. Hence, early IL-6 measurements in the emergency department can provide clinicians with an early picture of the prognosis of newly admitted patients, including injury severity and outcome, enabling better risk stratification. Larger-scale studies may help develop a preliminary injury severity and outcome score in the emergency department, especially when IL-6 is combined with other prognostic markers. An example of such an application is the TRACK-TBI cohort, in which the combination of IL-6 and other proinflammatory proteins discriminated among different severities of traumatic brain injury[29]. Overall, IL-6 is a promising indicator for outcome and extent of severity during the early stages of trauma care.

However, certain limitations to the feasibility of using IL-6 as a prognostic indicator in polytrauma must be considered. Notably, the magnitude of IL-6 response during inflammation is highly variable. Factors such as age, exercise, obesity, and medications can impact the circulating IL-6 levels[30]. Furthermore, there are also differences in techniques used to measure IL-6 levels[22]. In this meta-analysis, we observed substantial heterogeneity in the effect sizes for the correlation between IL-6 and ISS as well as the association between IL-6 and mortality.

Additionally, even within the literature, there is a high degree of variability in the IL-6 levels used to set cut-offs for predicting worse outcomes. For instance, Frink et al[26] reported a sensitivity of 16.7% in using IL-6 to predict MODS at a cut-off value of 761.7 pg/uL, while Rao et al[27] reported a sensitivity of 75% at a cut-off value of 106.56 pg. Before IL-6 can be used as a prognostic tool, studies are needed to standardize IL-6 measurement and determine the generalizability of the association across populations, such as the elderly and the obese.

Limitations

This meta-analysis has multiple limitations. First, data could not be pooled in a single step, since studies reported different types of correlation coefficients-Pearson and Spearman-which capture different association structures (linear vs monotonic) and have distinct statistical properties. To preserve analytical validity, we performed separate analyses for each type[31,32]. Finally, heterogeneity across studies in terms of trauma type, measurement timing, and study design may also have influenced the results.

There was wide variability in IL-6 levels measured within the first 24 hours across studies, with reported values ranging from 0.16 ± 0.06 pg/mL[14] to 569 pg/mL[18]. Furthermore, some studies did not report the mean or median IL-6 value. This variability and missing data made it difficult to directly compare studies or assess the relationship between IL-6 and ISS. This may have contributed to the high heterogeneity observed in the meta-analysis. Exploration of potential sources was limited by the small number of included studies, relatively homogeneous study design and quality, and incomplete reporting of methodological variables. Accordingly, the quantitative synthesis was based on summary data available in published reports. Another limitation is that, in some studies, IL-6 values for survivors and non-survivors were approximated from graphical representations rather than directly extracted from the text or tables. This introduces a risk of inaccuracy in the pooled estimates and could affect the reliability of the findings. The small number of studies also limits the reliability of Egger’s P-value for assessing publication bias from funnel plots. Finally, the observational nature of the included studies precludes causal inference, underscoring the need for additional RCT to confirm the observed associations. Consequently, pooled estimates should be interpreted cautiously and considered only hypothesis-generating.

CONCLUSION

This review provided evidence of a significant correlation between IL-6 levels measured within the first 24 hours after admission and ISS and mortality in polytrauma patients. IL-6 could, therefore, be a useful prognostic indicator of injury, severity, and survival. Future research should investigate whether early IL-6 levels, when combined with other biomarkers, can predict injury severity and adverse outcomes. Additionally, larger-scale studies should be conducted to assess the tool’s validity across diverse populations.