For Manuscript Submission, Check or Review Login please go to Submission Websites List.
For the academic login, please select your country in the dropdown list. You will be redirected to verify your credentials.
Predicting Post-Stroke Infections and Outcome with Blood-Based Immune and Stress MarkersMeisel A.a · Meisel C.b · Harms H.a · Hartmann O.c · Ulm L.a
aNeurocure Clinical Research Center, Department of Neurology and Center for Stroke Research Berlin, and bDepartment of Medical Immunology, Charité – Universitätsmedizin Berlin, Berlin, and cThermo Fisher Scientific BRAHMS GmbH, Hennigsdorf, Germany Corresponding Author
Department of Neurology, Charité – Universitätsmedizin Berlin
DE–10117 Berlin (Germany)
Tel. +49 30 450 560 026, E-Mail firstname.lastname@example.org
About one third of early deaths and poor outcomes after acute stroke are caused by potentially preventable stroke-associated complications, especially infections. Early identification of patients at high risk of infections and poor prognosis with biomarkers might help to initiate adequate therapies and guide treatment decisions. Acute injury of the central nervous system, including stroke, disturbs the normally well-balanced interplay between the sympathetic nervous system and the immune system, thereby impairing the antibacterial immune response in stroke patients. Changes in immune and stress markers, for example a reduction in HLA-DR expression on monocytes or an increase in serum catecholamine levels, occur very early after stroke onset, explain the high susceptibility of stroke patients to bacterial infections, and are predictive of infectious complications occurring up to 2 weeks after stroke. Outcome prediction is of utmost importance for decision-making in stroke units as well as in neurological intensive care units. However, to date the accuracy of outcome prediction by physicians and clinical scoring systems is only moderate. So far, only two blood-based biomarkers have been identified as independent predictors of outcome and mortality after stroke: the stress marker copeptin and midregional pro-atrial natriuretic peptide. Careful evaluation of prognostic markers is needed to prevent the occurrence of self-fulfilling prophecy.
© 2012 S. Karger AG, Basel
Stroke is one of the leading causes of severe long-term disability and death worldwide . Prognosis after stroke is highly variable and depends on patient-related factors, such as stroke severity and age, but also on potentially preventable stroke-associated complications, like increased intracranial pressure and infections [2,3], which account for one third of all in-hospital deaths after ischaemic stroke . Among post-stroke infections, bacterial pneumonia is of particular importance, as it causes almost 20% of in-hospital deaths after stroke [4,5,6,7] and has a negative impact on long-term outcome [3,8,9].
Underlying infections are the most common cause of fever after stroke . Pyrexia after stroke is independently associated with an increased risk of poor outcome, particularly in patients who have an early rise in body temperature after stroke . On the other hand, induced hypothermia is known to have neuroprotective properties and to improve outcome, for example after resuscitation . These observations imply that early cooling of patients might be effective in improving outcome after stroke. However, hypothermia is also known to suppress the immune system, and acute-stroke patients treated with hypothermia had increased rates of pneumonia . Up to now, cooling awake patients to 35°C after ischaemic stroke has been shown to be safe, but whether this improves functional outcome has not yet been tested in an adequately sized randomized clinical trial . Hence, the upcoming EuroHYP trial, which is a European, multicenter, open, randomized, phase III clinical trial, will assess the effects of therapeutic cooling in 1,500 awake adult patients with acute ischaemic stroke.
As post-stroke infections have been shown to be associated with poor outcome in animal models of stroke and in human trials [2,15], the question arises whether preventive antibiotic treatment may not only reduce the incidence of infections, but also improve overall outcome. Indeed, in a preclinical stroke model, prophylactic antibiotic treatment prevented infections and improved outcome . A recent meta-analysis of all published trials investigating the effects of prophylactic antibiotic treatment in acute-stroke patients concluded that this treatment strategy is effective in preventing post-stroke infections, but does not reduce mortality . However, as all of these clinical trials were not sufficiently powered to investigate the impact of a preventive antibiotic treatment on outcome, the observed effects need to be evaluated in large trials that include patients at risk of post-stroke infections. Table 1 summarizes recently finalized or ongoing clinical trials aimed at treating post-stroke infections by antibiotic therapy.
A major drawback of preventive antibacterial therapy is the potential promotion of antibiotic resistance in common bacteria, since increased use of antibiotics can increase resistance rates. Biomarkers might help to select patients who need antibiotic therapy (fig. 1), thereby lowering the risk of increased resistance rates. Furthermore, they might provide information on a patient’s prognosis early after stroke, which might be important to guide treatment decisions. In a patient who is expected to have a good outcome, early treatment of complications (e.g. infections) might be desirable to ensure his or her good probability of improvement. On the other hand, for patients in whom a reduction of therapeutic efforts is considered, reliable prognostic information would be helpful to assist families and physicians in their difficult decision.
Biomarkers are defined as biological, biochemical, or biophysical parameters that can be monitored objectively and reproducibly in humans or animal models . Several categories of biomarkers have been studied with regard to their power to predict complications and outcome after stroke, e.g. physical markers like body temperature or blood glucose [19,20], imaging markers including perfusion- and diffusion-weighted imaging (PWI/DWI) of lesion volumes , and blood or cerebrospinal fluid markers. Among the latter, immune and inflammatory markers and proteins involved in cerebral and cardiac tissue damage or haemostasis are under investigation . The levels of the neuronal and glial damage markers neuron-specific enolase (NSE) and S100-β, for example, have been demonstrated to correlate with infarct volume . Moreover, high levels of fibrinogen have been associated with poor outcome 3 months after stroke . In the following review, we will focus on blood-based immune and stress markers predicting outcome and infections after stroke and their implications for clinical care.
Even in specialized stroke units, post-stroke infections remain one of the main complications in acute stroke, with frequencies between 21 and 65% . A number of predisposing factors are considered to account for the high incidence of bacterial pneumonia following stroke, including several factors known to facilitate aspiration, such as impaired protective reflexes, dysphagia and mechanical ventilation [4,6,25]. In addition, experimental and clinical evidence suggests that acute central nervous system injury, including stroke, directly impairs antibacterial host defence, thereby increasing susceptibility to infections [26,27,28,29]. In particular, down-regulation of systemic cellular immune responses, i.e. a rapid numerical decrease in peripheral blood lymphocyte subpopulations and functional deactivation of monocytes (fig. 2), T helper and invariant natural killer T cells , has been reported. Furthermore, signs of immunodepression are more prominent and recover more slowly in patients who develop infectious complications [27,31,32,33,34,35]. Importantly, reduced monocytic HLA-DR expression on day 1 after stroke onset was a strong independent predictor of subsequent post-stroke infections in the PANTHERIS trial . In this trial, infections occurred between day 2 and day 7 after stroke onset; on average, both pneumonia and urinary tract infections were diagnosed on day 5. Moreover, changes in immune markers on day 1 after stroke, like reduced monocytic HLA-DR expression or low CD4+ T cell counts, independently predicted subsequent post-stroke infections developing within 2 weeks after stroke [31,35].
Acute central nervous system injury, including stroke or traumatic brain injury, disturbs the normally well-balanced interplay between the nervous and the immune system by compromising sympathetic and parasympathetic neural connections with lymphoid organs and humoural components, including the hypothalamo-pituitary-adrenal axis [2,36,37]. A key role of the sympathetic nervous system in impaired antibacterial immune response after stroke has been shown in a mouse model of focal cerebral ischaemia: in that model, impaired lymphocyte responses and increased susceptibility to bacterial infections could be effectively prevented by blocking the sympathetic nervous system [15,30,38]. Moreover, Chamorro et al.  demonstrated that the development of infections early after acute ischaemic stroke is associated with enhanced activation of the sympathetic adrenomedullar pathway as indicated by significantly increased plasma catecholamine levels on day 1 after stroke in patients who subsequently (days 2–7) developed infections . Similarly, in the PANTHERIS trial , patients with infections in the placebo group had significantly higher urine norepinephrine levels on day 1 and 2 after stroke compared to patients without infections . These clinical findings support the concept of a stress-mediated immunodepression driven by the sympathetic nervous system and hypothalamo-pituitary axis as an essential facilitating factor for the development of post-stroke infections.
These findings might also have implications for the treatment of post-stroke infections. Selective immunomodulation to prevent post-stroke infections may be an alternative to preventive antibiotic treatment . For example, administration of α-galactosylceramide, an activator of invariant natural killer T cells, increased systemic interferon-γ (IFN-γ) levels, and reduced bacterial infections and lung damage after experimental stroke in mice . Likewise, systemic application of IFN-γ prevented post-stroke infections in an experimental stroke model . To our knowledge, only one drug with known immunomodulatory properties, i.e. granulocyte colony-stimulating factor (G-CSF), has been investigated in clinical trials of acute stroke treatment. G-CSF is a haematopoietic cytokine responsible for the mobilization and differentiation of haematopoietic stem cells. It is in widespread clinical use to mobilize stem cells for transplantation in patients with haematological malignancies or for the treatment of chemotherapy-associated neutropenia. In addition to its effects on the haematopoietic system, G-CSF has also been shown to have neuroprotective and neurotrophic properties in experimental stroke models . In the multicenter, randomized, placebo-controlled phase IIa trial AXIS, the safety and tolerability of intravenous G-CSF treatment have been tested in 44 acute ischaemic stroke patients. G-CSF led to a significant increase in neutrophils and monocytes . The subsequent randomized, double-blind AXIS 2 trial, which included 328 stroke patients, failed to demonstrate an improvement of 3-month outcome by G-CSF treatment (press release of SYGNIS Pharma AG, December 15, 2011). The effects on post-stroke infections were reported in neither trial.
In summary, changes in immune and stress markers occur very early after stroke onset, explain the high susceptibility of stroke patients to bacterial infections, and are able to predict infectious complications occurring within 2 weeks after stroke (table 2). Further research is needed to better understand the mechanisms of brain-immune interaction after acute brain injury, and the long-term consequences immunomodulatory therapies may have for stroke outcome .
The National Institutes of Health Stroke Scale (NIHSS) is a quantitative assessment of stroke-related neurological deficits. It can be performed quickly  and with high reliability and validity . Originally developed for use in acute-stroke therapy trials [43,45], it has been shown to have predictive capacity concerning functional outcome up to 6 months after stroke . Several blood-based inflammation markers, like C-reactive protein [47,48], high-sensitive C-reactive protein [49,50], white blood cell count , monocyte chemotactic protein-1  or interleukin-6 , have also been associated with outcome, but all of them failed to improve the prognostic capacity of the NIHSS, which limits their clinical relevance. A role of the innate immune system in outcome prediction has recently been demonstrated: the monocyte subtype CD14highCD16 and higher expression of Toll-like receptor 4 in monocytes were independently associated with poor prognosis in acute stroke patients [53,54], and deficiency of mannose-binding lectin, a complement activator, was independently associated with better outcome in experimental and human stroke .
Stress-responsive cytokines are induced after stroke and might predict outcome. For example, elevated serum concentrations of growth differentiation factor 15 are associated with an unfavourable outcome after ischaemic stroke . To date, there are only two stress markers that independently predict functional outcome and mortality: copeptin, a fragment of provasopressin, and midregional pro-atrial natriuretic peptide (MR-proANP). Blood levels of the stress marker copeptin upon admission (0–72 h after symptom onset) predicted functional outcome and mortality within 90 days similarly to the NIHSS (c-index of 0.82 for mortality and 0.73 for outcome). In multivariate logistic regression analysis including known outcome predictors such as the NIHSS, age and female sex, copeptin remained an independent predictor of an unfavourable functional outcome and mortality . Even in a 1-year follow-up of the same patient cohort, the initial copeptin concentrations independently predicted functional outcome and death . MR-proANP can be seen as an indirect stress marker, as higher levels of natriuretic peptides in stroke patients are associated with increased sympathetic activation . Elevation of MR-proANP upon admission (0–72 h after symptom onset) was strongly associated with mortality and poor functional outcome at day 90 after stroke (c-index of 0.86 for mortality and 0.70 for outcome). Models incorporating copeptin or MR-pro-ANP and the NIHSS even improved their accuracy in the prediction of an unfavourable outcome or mortality [57,59]. The predictive capacity of copeptin and MR-proANP was further corroborated by the PANTHERIS trial : Copeptin and MR-proANP levels 3 days after stroke showed high sensitivity and specificity in the prediction of poor outcome (death or Barthel Index <20) 90 days after stroke (c-index 0.83 and 0.79, respectively).
However, some blood-based biomarkers, such as interleukin-6 or N-terminal pro-brain natriuretic peptide, might not have sufficient predictive power to be of clinical use to predict poor outcome after stroke [60,61]. A sensitive and specific prediction of outcome is of paramount importance in the management of critically ill patients, in whom decisions about life-sustaining therapies are made. These include decisions not to institute therapies that would otherwise be warranted (withholding) and decisions about the discontinuation of treatments that had been started (withdrawing). Withholding and withdrawing treatments are considered legally and ethically equivalent , although this is in conflict with social or psychological convictions that they are different acts  and they might have different implications for clinical care, e.g. regarding the importance of consent  or their potential impact on patients’ outcome .
Decisions to forgo treatment in intensive care units have been shown to be associated with several factors, including higher age of the patients, the severity of their illness, pre-existing severe medical conditions, physicians’ and patients’ religious affiliation, patients’ wishes, and the physicians’ prediction of a low likelihood of survival and poor future functional outcome [64,65,66,67,68]. Especially the latter is a strong determinant of life support limitation. Outcome prediction (chances of recovery, expected time course of recovery and anticipated quality of life) influences treatment decisions of physicians [64,69,70,71] and is the most important criterion for patients and their relatives to make a decision of treatment withdrawal [67,72] – although the accuracy of early outcome prediction by physicians or currently used scoring systems, e.g. the Acute Physiology and Chronic Health Evaluation Score (APACHE) II, is only moderate . Against this background, the development of sensitive and specific biomarkers to improve outcome prediction is of great importance for clinical practice and an important research field. A number of genes involved in immune regulation and inflammatory responses show characteristic changes in their expression pattern after stroke; these changes precede infections, and a bad functional outcome and death (fig. 3, 4). However, before turning these immune markers into useful tools for clinical practice, some questions still have to be answered. First, one has to be sure that changes in inflammation markers used to predict outcome are independent of the occurrence of complications, e.g. infections that influence the outcome negatively . Furthermore, cut-off values of biomarkers are presumably affected by multiple parameters, as is seen with neuron specific enolase (NSE). In comatose survivors after cardiopulmonary resuscitation, NSE levels exceeding 33 µg/l were an established predictor of poor outcome [75,76]. With the introduction of the novel treatment approach of therapeutic hypothermia, the previously validated cut-off levels were not useful anymore, as survivors undergoing hypothermia treatment had markedly higher NSE cut-off levels for a bad outcome compared to non-hypothermia patients [77,78].
A reduction in the number of therapeutic procedures based on biomarkers that predict a bad outcome has to be evaluated carefully to prevent the creation of self-fulfilling prophecy. If, for example, antimicrobial therapy is withheld in patients who are considered to be at high risk of having a bad outcome, this may lead to a higher probability of bad outcomes in these patients – whether or not the initial prediction was correct. But as the rate of bad outcomes will then be higher in the group which was expected to have a bad outcome, the prediction will be believed to be true.
In conclusion, blood-based immune and stress markers might identify patients at high risk of post-stroke infections as well as patients with unfavourable outcomes. Thus, biomarker guidance is a promising strategy to tailor preventive treatment of post-stroke infections, thereby improving long-term outcomes. However, further clinical studies for the characterization of prognostic markers are needed to provide clinicians with reliable information.
We thank Stephanie Ohlraun for her valuable comments on the manuscript. This work was supported by the German Research Foundation (Exc 257), the Federal Ministry of Education and Research (01 EO 08 01), the Helmholtz Association (SO-022NG) and has received funding from the European Community’s Seventh Framework Programme (FP7/2007–2013) under grant agreement No. 201024 (all given to A.M.).
C.M. and H.H. have received speaker’s honoraria from Bayer Vital GmbH. A.M. has received speaker’s honoraria from Bayer Vital GmbH and Wyeth Pharma GmbH. A patent application on anti-infective agents and immunomodulators used for preventive therapy following an acute cerebrovascular accident has been filed to the European Patent Office (PCT/EP03/02246) (patent owner Charité – Universitätsmedizin Berlin, inventor A.M., C.M.). O.H. is employed as a biostatistician by Thermo Fisher Scientific Brahms GmbH, Germany.
Department of Neurology, Charité – Universitätsmedizin Berlin
DE–10117 Berlin (Germany)
Tel. +49 30 450 560 026, E-Mail email@example.com
Copyright: All rights reserved. No part of this publication may be translated into other languages, reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, microcopying, or by any information storage and retrieval system, without permission in writing from the publisher or, in the case of photocopying, direct payment of a specified fee to the Copyright Clearance Center.
Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug.
Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.