IL-6 is a predictive biomarker for stroke associated infection and future mortality in the elderly after an ischemic stroke

Highlights

• Infection after stroke in the elderly causes poor recovery and often death.

• There is no good way of predicting who will have a stroke associated infection.

• IL-6 after an ischemic stroke independently predicts future infection.

• IL-6 predicts increased risk of death within the first 2 years after a stroke.

• Therefore, using IL-6 as a biomarker may determine who to treat with antibiotics.

Abstract

Background and purpose

Stroke associated infection (within the first seven days) occurs in approximately half of stroke patients and is associated with a worse prognosis, especially in the elderly. It is uncertain what factors predict stroke associated infection, yet identification of a suitable biomarker for infection may allow early and appropriate intervention with antibiotics. The aims of this study were to: a) identify independent risk factors for stroke associated infection, and b) test relationships between these risk factors and mortality at 2 years.

Methods

Eight-two elderly patients were assessed within 72 h of stroke. Data on stroke severity (Barthel Index), stroke associated infection and mortality at 2 years were collected. Inflammatory biomarkers at baseline and 6 months were measured by ELISA. Logistic regression was used to identify risk factors for stroke associated infection and death.

Results

Patients with stroke associated infection, especially pneumonia, had increased IL-6, more severe strokes, and higher mortality. IL-6 was independently associated with stroke associated infection (OR = 19.2, [95%CI 3.68, 100], p < 0.001), after adjustment for other risk factors and cytokines. IL-6 was also independently associated with 2 year mortality (OR = 9.2, [1.0, 85.1], p = 0.031).

Conclusions

These data suggest that IL-6 may be a key biomarker for predicting stroke associated infection and mortality in the first two years post stroke.

Introduction

The main objective of this study was to identify one or more biomarkers to enable the early prediction of stroke associated infection. The effects of stroke-associated infections (SAI) on short-term outcome are well known, increasing the risk of early death and institutionalization, especially in the elderly. For example, patients with a SAI were three times more likely to die within the first 5 days compared to those who did not have a SAI (Kwan and Hand, 2007). The effect of SAI on long term outcome was demonstrated more recently in a prospective study which showed that the risk of dying over 3 years was significantly higher in patients who had a SAI in the first 2 weeks after a stroke (Kwan et al., 2013). Despite the detrimental effects of SAI, there is no easy way of predicting which patients will develop an infection, or which patients are at increased risk of mortality. Previous studies have demonstrated several parameters associated with SAI, including stroke severity and use of nasal gastric tubes. However, if a simple, quick and inexpensive test to measure one or more biomarkers could be developed to identify those patients most at risk, this may allow early and appropriate intervention with antibiotics. The effect of prophylactic antibiotic use in SAI is currently under investigation (Harms et al., 2008, Nederkoorn et al., 2011, Vargas et al., 2006) and therefore identification of appropriate biomarkers is particularly timely.

Acute stroke initiates an inflammatory response in the brain and peripherally, whilst response to infection also involves inflammation and the acute phase response. Recently there has been evidence that stroke results in suppression of the immune system which may also predispose towards infection (Braun et al., 2007, Chamorro et al., 2006, Chamorro et al., 2007a, Chamorro et al., 2007b, Emsley and Hopkins, 2008, Haeusler et al., 2008, Klehmet et al., 2009, Prass et al., 2003, Vogelgesang et al., 2008). There are several candidate biomarkers that are worth exploring as risk factors for SAI. Proinflammatory cytokines may be early biomarkers for infection, as well as having plausible biological actions that predispose towards infection. For example, IL-6 is an important acute phase mediator that is elevated during infection (Smith et al., 2006), with multiple effects that may be important. Plasma IL-6 levels are elevated in acute stroke, and higher levels of IL-6 are associated with larger infarct volumes and poor outcome (Smith et al., 2004, Smith et al., 2006, Whiteley et al., 2012). TNFα is also a good candidate biomarker as there is increasing evidence that inflammation may suppress both innate and adaptive cellular immunity in stroke via TNFα (Haeusler et al., 2008, Urra et al., 2009) which suggests that a proinflammatory state may increase risk of infection. A third candidate cytokine is IL-1β which has multiple effects on cytokine and acute phase protein gene expression, as well as both innate and adaptive immunity via lymphocyte stimulation and an involvement in clearance of bacterial infections (Miller et al., 2007). These novel biomarkers may be used in combination with more traditional parameters, such as blood pressure and age, to produce a sensitive and predictive model for SAI.

Therefore, given the link between SAI and poor outcome, the current concern over inappropriate antibiotic use, and a lack of a robust predictive model for SAIs, the aims of this study were to: a) identify independent risk factors for SAI and b) test relationships between these risk factors and mortality at 2 years.