In this preliminary study we have shown that there is a clear demonstrable cytokine response during the early stages following major trauma. We have also shown, for the first time in trauma patients, that the early plasma cytokine concentrations of two prototypical cytokines involved in these inflammatory processes (IL-6 and IL-10) are significantly greater in patients who then proceed develop clinical phenotypes of worsening organ dysfunction and SOFA scores. Early identification of high-risk patients provides a potential window during which interventions aimed at preventing the evolution of organ failure (invasive monitoring, close nursing care etc) may be commenced to those most at risk and thereby optimise the utilisation of scarce resources. It is recognised that these cytokines render themselves well for point of care testing and thereby are ideal targets for bedside diagnostic tests that may be built into risk stratification algorithms. Despite there being statistically significant differences between groups that develop organ dysfunction and those who do not, our study shows a wide scatter in the distribution of the concentrations of both cytokines implying that multiple factors may be involved in the regulation of these responses. It is therefore likely that even though a biomarker-based approach to risk stratification may be used to improve the accuracy of clinical decision-making, they are unlikely to replace a traditional, clinician lead, multi-dimensional assessment of a trauma victims.
In the study we have used mortality and a SOFA score of ≥ 2 on day 7 as indicative of an adverse clinical outcome. In established ICU patients it is known that SOFA scores have a linear relationship with ICU mortality (19). As the score represents a composite score derived across six domains - each of which may be scored between 0 (normal) to 4 (severe organ failure), the maximum possible score is 24 (20). Therefore, the threshold of ≥ 2 represents the lower end of the spectrum involving mild dysfunction affecting one or more organ systems. This threshold on day 7 was selected by the study team as the main outcome measure, as the study was aimed at identifying patients who were clinically deteriorating, but still amenable to clinical interventions. Many of the patients categorised into the adverse outcome group therefore survived their injury and were successfully discharged from hospital. The fact that they were in an intensive care unit is likely to have contributed to the timely recognition and reversal resulting in the favourable outcomes. The finding that even at such a low threshold, there were significant differences in cytokine profiles is encouraging when considering the potential value of these two cytokines as suitable targets for developing point of care tests. This low threshold also could potentially explain why our data shows a relatively wide scatter (represented by the wide inter-quartile intervals) and overlap between groups despite reaching statistical significance. It is therefore possible that by setting the threshold higher to capture patients who are more unwell it may be possible to reduce this overlap, thus better demonstrating the potential value of these cytokines in risk stratification.
The role of cytokines in the evolution of organ dysfunction has been studied extensively in patients with sepsis. Comparatively similar longitudinal studies in trauma patients are few and far between. Whilst many different cytokines have been previously examined following traumatic injury (15, 21–23), there is limited further research to corroborate these findings. Of all cytokines previously examined, IL-6 and IL-10 have been highlighted to be of great potential in determining injury severity following trauma. Multiple studies have shown both IL-6 (13, 15, 24) and IL-10 (24, 25) concentrations to be aberrantly elevated in the period following trauma. However, limited examination into their diagnostic or prognostic potential for long term traumatic injury outcome and organ dysfunction has been conducted (24, 26, 27). Our study is unique as we have undertaken repeated measurements of these two key cytokine profiles and hence have determined the temporal evolution of these profiles between two distinct clinically relevant cohorts.
An extensive body of research reported previously has also highlighted the potential of IL-6 in particular as a biomarker of traumatic injury severity (13, 24, 26, 28–30) and clinically important outcomes such as; multi-organ dysfunction (24, 25, 27, 31, 32), sepsis (33) and mortality (13, 24). A study by Frink et al. (32) showed that IL-6 was found to be significantly higher in patients with MODS compared to those without organ dysfunction in 143 trauma patients. Additionally, they identified that early plasma IL-6 concentrations could identify patients who were at higher risk of developing MODS with high specificity but low sensitivity (32). The concentrations of IL-6 in our study were significantly greater in groups with worsening organ functions on day 7 (Fig. 1) and day 5 (Supplementary Fig. 3) and in patients who were showing worsening organ functions between days one and five (Fig. 4A and 4B). IL-6/IL-10 ratio is considered to be an index showing the balance between pro and anti-inflammatory pathways (24, 29). In our analyses the difference in this ratio reached statistical significance only in the ΔSOFA outcome groups on day 5 (Fig. 4E and 4F) after excluding 2–4 outliers. These observed differences were not evident against the ΔSOFA outcome groups on day 7 (Fig. 3). These discrepancies are attributable to the multitude of factors involved in the regulation of particularly IL-10 – a key counter regulatory cytokine which is produced by the body to quench the pro-inflammatory pathways activated by trauma (34). Unlike IL-6 levels which were significantly different on day 3 itself (Fig. 1), difference in IL-10 concentrations between the two outcome groups reached statistical significance only on day 5. This observation supports the role of IL-10 as a counter regulatory cytokine, with the relevant plasma concentrations lagging behind that of IL-6 which mediates the pro-inflammatory responses that trigger IL-10 production. Compared to IL-6 levels, the decline in IL-10 concentrations did not follow a single unimodal decline (Fig. 1) suggesting further that the regulation of IL-10 is more complex and hence plasma concentrations are potentially less predictable. IL-10 (and other counter regulatory cytokines) potentially contribute to immunosuppression thereby making these patients more vulnerable to hospital acquired infections at this stage of their illness. The finding of significantly greater concentration of plasma [lactate] on day 5 (Fig. 2), a common association with sepsis, is a very pertinent and interesting finding in this context. The production and plasma concentrations of counter regulatory cytokines such as IL-10 are therefore subject to multiple factors including fluid resuscitation, antibiotics, secondary infections, and other supportive care. The cumulative impact of these therapeutic interventions are likely to be manifest by day 7 and hence, in the present study we have explored the relationships between plasma [cytokine] on clinical outcomes at day 5 and day 7. Figure 4 shows that IL-6 and IL-6/IL-10 ratios were significantly greater in the group showing deterioration in the SOFA score between days 1 and 5. These differences are not evident at ΔSOFA Day 7 as the effects of interventions become more manifest making IL-10 concentrations more variable. Our study further develops the existing body of data in this area by showing that the early absolute values as well as the temporal changes in IL-6 and IL-10 during the first few days (days 1 and 3) after injury may be an early marker of organ dysfunction (Figs. 3 and 4). More importantly these differences are evident early in the course of the illness (days 1–3) where potential interventions aimed at reversing these processes are possible.
Our study did not show any differences in CRP or Plasma [lactate] between the two outcome groups on days 1–3. While no significant differences were seen in CRP on day 5, plasma [lactate] was higher in the group developing worsening organ functions (Fig. 2). Serum [Lactate] in particular has been highlighted as a useful predictor of clinical outcome in trauma patients (35–38). These studies however were almost uniformly undertaken in the emergency departments and hence would have been a marker of severity of shock and the quality of care in the pre-hospital environment. Our patients were recruited after their initial management in an A&E department where they would have received adequate resuscitation and reversal of shock/tissue hypo perfusion. It is therefore not surprising that there was a poor association between serum [lactate] on days 1 and 3 and organ dysfunction on day 7 (Fig. 2). Serum [lactate] levels on day 5 on the other hand was significantly greater (Fig. 2), perhaps with the onset of sepsis or other factors that may affect tissue perfusion/metabolism, in the group showing greater organ dysfunction. CRP, usually a marker of inflammation or sepsis was not significantly different between the two groups at any point in our study.
An important limitation of the present study needs emphasis. As this was an observational study, the effects of treatment, remains an important confounding factor. Secondary infections, use of antibiotics, fluid management and ventilation associated lung injury (VILI) are all examples of potential modifiers of early cytokine responses and may account for the relatively wide confidence intervals in our data and considerable overlap between groups. Despite these confounders, we have shown that significant differences do exist between cohorts of patients who undergo worsening of organ functions or not. Larger longitudinal studies are required to determine the value of these early cytokine responses in predicting clinical outcomes and if their incorporation into decision-making algorithms is a viable option for improved risk stratification.