Prediction Of Individual Outcome Research Articles

GFAP and S100B in the acute phase of mild traumatic brain injury

The biomarkers glial fibrillary acid protein (GFAP) and S100B are increasingly used as prognostic tools in severe traumatic brain injury (TBI). Data for mild TBI are scarce. This study aims to analyze the predictive value of GFAP and S100B for outcome in mild TBI and the relation with imaging. In 94 patients biomarkers were determined directly after admission. Collected data included injury severity, patient characteristics, admission CT, and MRI 3 months postinjury. Six months postinjury outcome was determined with Glasgow Outcome Scale Extended (GOSE) and return to work (RTW). Mean GFAP was 0.25 μg/L (SD 1.08) and S100B 0.54 μg/L (SD 1.18). In 63% GFAP was not discernible. GFAP was increased in patients with an abnormal CT (1.20 μg/L, SD 2.65) compared to normal CT (0.05 μg/L, SD 0.17, p < 0.05). Also in patients with axonal injury on MRI GFAP was higher (0.65 μg/L, SD 0.91 vs 0.07 μg/L, SD 0.2, p < 0.05). GFAP was increased in patients with incomplete RTW compared to complete RTW (0.69 μg/L, SD 2.11 vs 0.12 μg/L, SD 0.38, p < 0.05). S100B was not related to outcome or imaging studies. In multivariate analysis GFAP was not predictive for outcome determined by GOSE and RTW. A relation between GFAP with imaging studies and outcome (determined by RTW) was found in contrast to S100B. As the positive predictive value of GFAP is limited in this category of TBI patients, this biomarker is not suitable for prediction of individual patient outcome.

Molecular classification of breast carcinoma

Breast cancer is a complex disease encompassing multiple tumour entities each with a characteristic morphology and behaviour. Current clinical practice relies on the recognition of various pathology prognostic factors to guide patient management, including histological type and grade, stage and biomarker receptor status. However, there is increasing concern that these parameters are of limited value for the accurate prediction of individual patient outcome. The introduction of genome-wide microarray-based expression profiling studies has allowed better understanding of the molecular underpinning of several characteristics of breast cancer, including histological grade and metastatic potential. Expression profiling has also facilitated the identification of prognostic and predictive gene expression signatures and novel therapeutic targets. Here we review the evolution of molecular classification of breast cancer, including special types, the implications for clinical management, limitations of findings thus far and predictions for the future.

Is the Metabolic Syndrome Associated with Cardiovascular Disease, Diabetes Mellitus, and Increased Physical Limitations in an Elderly Population if Multiple Chronic Illnesses are Taken into Consideration?

The concept of metabolic syndrome (MetS) was initially designed for long-term outcome prediction in middle aged or younger individuals, but it is unclear whether MetS can be applied in older people. We explored the clinical utility of MetS for prediction of various outcomes in older people. First, we investigated whether MetS is associated with CVD and DM in the short term. Second, we took multiple chronic illnesses into consideration to examine whether MetS is associated with increased physical limitations. Participants were selected from the “Social Environment and Biomarkers of Aging Study, 2000” and the “2003 Survey of Health and Living Status of the Middle-Aged and Elderly in Taiwan” aged ≥65 years for whom complete data were available. A total of 543 older Taiwanese were included. For older people who had MetS compared to those without the syndrome in 2000, the risk of CVD was 1.94 times higher in 2003 ( p < 0.01) and that of DM was 3.02 times higher in 2003 ( p < 0.01) after controlling for age, sex, and chronic illnesses. Hierarchical regression models of physical functions showed that MetS did not increase the explained variance in physical functional limitations. MetS has short-term (3 years) effects on CVD and DM. After taking chronic illnesses into consideration, no association between MetS and physical functioning was obvious. More studies are needed to clarify the relationship among MetS, chronic illnesses, and physical functioning.

Chromium-Picolinate Therapy in Diabetes Care: Molecular and Subcellular Profiling Revealed a Necessity for Individual Outcome Prediction, Personalised Treatment Algorithms &amp; New Guidelines

Global figures clearly demonstrate inadequacy of current diabetes care: every 10 seconds one patient dies of diabetes-related pathologies. Nephropathy is the leading secondary complication of the disease. Nutritional supplement by chromium-picolinate is assumed to have beneficial therapeutic effects. However, potential toxic effects reported increase concerns about safety of chromium-picolinate. The experimental design aimed at determining, whether the treatment with clinically relevant doses of chromium-picolinate can harm through DNA damage and extensive alterations in central detoxification / cell-cycle regulating pathways in treatment of diabetes. Well-acknowledged animal model of db/db-mice and clinically relevant doses of chromium-picolinate were used. As an index of DNA-damage, measurement of DNA-breaks was performed using "Comet Assay"-analysis. Individual and group-specific expression patterns of SOD-1 and P53 were evaluated to give a clue about central detoxification and cell-cycle regulating pathways under treatment conditions. The study was performed in a double-blind manner. Experimental data revealed highly individual reaction under treatment conditions. However, group-specific patterns were monitored: highest amount of damaged DNA--under the longest treatment with high doses, in contrast to groups with low doses of chromium-picolinate. Comet patterns were intermediate between untreated diabetised and control animals. Expression patterns demonstrated a correlation with subcellular imaging and dosage-dependent suppression under chromium-picolinate treatment. This article highlights possible risks for individual long-term effects, when chromium-picolinate is used freely as a therapeutic nutritional modality agent without application of advanced diagnostic tools to predict risks and individual outcomes. Targeted measures require a creation of new guidelines for advanced Diabetes care.

Biology of bone, cartilage and connectve tissue disease: 92. WNT3A Modulates Chondrocyte Phenotype through Activation of both Canonical and Non-Canonical Pathways

Background: Recent data indicate that Wnt signalling is a major player in the development of osteoarthritis and cartilage breakdown and therefore point to Wnt signalling as an extremely attractive target for osteoarthritis. However some experiments have given contrasting results in the way that both excessive activation and inhibition of this pathway result in increased cartilage breakdown. Wnts can signal through multiple pathways one of which involving the cytoplasmic molecule β-catenin and another one Ca2+ signalling and CaMKII. We have hypothesized that cartilage breakdown following Wnt stimulation is underpinned by a different molecular mechanism of that induced by Wnt inhibition. To test this hypothesis we have blocked at different levels individual pathways mediating Wnt signalling in chondrocytes and have developed a model demonstrating for the first time that one individual Wnt, ligand, Wnt3a, can simultaneously signals through multiple pathways with distinct outcomes and that the individual pathways negatively regulate each other. Methods: Primary articular chondrocytes were isolated from preserved cartilage areas of patients undergoing knee arthroplasty for osteoarthritis. Treatments with Wnt3a alone or in combination with DKK1 (inhibitor of the β-catenin-dependent pathway) or KN93 (CaMKII inhibitor) were performed on confluent primary monolayers or in vivo in the ectopic cartilage formation assay. The activation of the β-catenin dependent pathway was evaluated by reporter assay, nuclear accumulation of β-catenin, or expression of Axin-2. The activation of the Ca2+/ CamKII-dependent pathway was evaluated by Ca2+ mobilization or CamKII phosphorylation (immunofluorescence). Gene expression analysis of proliferation and phenotype markers such as PCNA, COL2A1, Aggrecan and SOX9 was performed by quantitative real time PCR. Results: Wnt3a induced the activation of both the β-catenin-dependent and the CaMKII-dependent pathways. More importantly we have separated the effects that are directly dependent on the activation of the β-catenin-dependent pathway, including proliferation, from those that are CaMKII-dependent such as disruption of the chondrocyte phenotype as evaluated by downregulation of differentiation markers. Most importantly we have built up a model that allows prediction and targeting of individual outcomes of Wnt signalling and explains the paradoxical results of the previous literature. Conclusions: Our findings will allow to therapeutically target individual outcomes of Wnt signalling, for instance to prevent chondrocyte de-differentiation without affecting crucial anabolic processes such as cell proliferation. It will also allow exploition of this extremely potent signalling pathway for therapeutic use in an unprecedented specific manner thereby avoiding potential undesired effects. Disclosure statement: The authors have declared no conflicts of interest.

Prediction of Individual Long-term Outcomes in Smoking Cessation Trials Using Frailty Models

In smoking cessation clinical trials, subjects commonly receive treatment and report daily cigarette consumption over a period of several weeks. Although the outcome at the end of this period is an important indicator of treatment success, substantial uncertainty remains on how an individual's smoking behavior will evolve over time. Therefore it is of interest to predict long-term smoking cessation success based on short-term clinical observations. We develop a Bayesian method for prediction, based on a cure-mixture frailty model we proposed earlier, that describes the process of transition between abstinence and smoking. Specifically we propose a two-stage prediction algorithm that first uses importance sampling to generate subject-specific frailties from their posterior distributions conditional on the observed data, then samples predicted future smoking behavior trajectories from the estimated model parameters and sampled frailties. We apply the method to data from two randomized smoking cessation trials comparing bupropion to placebo. Comparisons of actual smoking status at one year with predictions from our model and from a variety of empirical methods suggest that our method gives excellent predictions.

Predictive value of B-type natriuretic peptide levels in patients with paroxysmal and persistent atrial fibrillation undergoing pulmonary vein isolation

Atrial fibrillation (AF) is the most common cardiac arrhythmia reducing the health-related quality of life. Radiofrequency catheter ablation (CA) became the therapy of choice in patients with drug-resistant AF with success rates between 30% and 86%. However, CA remains a challenging intervention with major complications in about 5% of cases. Therefore, stratification into high and low success patient groups would be helpful. The aim of this study was to investigate the predictive value of B-type natriuretic peptide (BNP) on the outcome of pulmonary vein isolation (PVI) in patients with paroxysmal (PAF) and persistent (Pers-AF) atrial fibrillation. In 73 patients (median age 53years, 77% men) undergoing PVI for drug-refractory PAF (n = 45) or Pers-AF (n = 28), the serum BNP concentration was measured before and 3months after the ablation procedure to assess any association of pre- and post-interventional BNP concentrations with therapeutic outcome. The patients had suffered from AF for a median of 40months. No patient had structural heart disease or an impaired left-ventricular ejection fraction. A total of 54 patients (74%) had stable sinus rhythm 3months after PVI. The median baseline BNP levels in both PAF and Pers-AF patients were significantly lower in patients with a 3-month successful PVI than those in which it was unsuccessful, 57.5pg/ml (20.4-87.9) versus 159.0pg/ml (124.1-177.5; p = 0.001) in PAF patients and 90.3pg/ml (41.0-155.0) versus 176pg/ml (89.6-297.4; p = 0.026) in patients with Pers-AF, respectively. A multiple logistic regression analysis identified pre-interventional BNP levels as the only independent predictor for 3-month PVI outcomes (p = 0.010). Nevertheless, in this study, the predictive value of BNP for PVI outcomes was not high enough to permit individual outcome prediction. After successful PVI, BNP levels were significantly lower in patients with PAF and Pers-AF (median changes -16.9 and -23.8pg/ml; p = 0.010 and p = 0.022, respectively), but not in patients with AF in follow-up (median change 9.0pg/ml and -29.6pg/ml; p = 1.000 and p = 0.109, respectively). Pre-ablation BNP level seems to be an independent marker for successful PVI procedures in patients with paroxysmal and persistent AF; however, the observed level of association is moderate.

AED discontinuation may not be dangerous in seizure-free patients

The decision to continue or to stop anticonvulsant treatment in patients with prolonged seizure remission is still a matter of debate. There is no evidence that continued treatment guarantees permanent seizure freedom. Up to 50% of patients achieve terminal remission (i.e., off drugs). Almost one-half of patients with untreated epilepsy are seizure-free for more than 5 years and the number of those with continuing seizures tends to decrease over time. In contrast, a substantial fraction of patients on chronic treatment with antiepileptic drugs (AEDs) may be exposed to clinically relevant adverse effects. Other disadvantages of continuing treatment indefinitely include the teratogenic risk, drug interactions, and the concern that AEDs may be unnecessary. The probability of remaining seizure-free after treatment discontinuation is about 70% at 2 years. Although different combinations of factors are thought to affect the risk of relapse, the prediction of individual outcome before withdrawal remains uncertain. As a consequence, the decision to withdraw or withhold treatment must be still individualized. In any patient, the decision to discontinue treatment should also take into account social aspects like driving license, job and leisure activities as well as emotional and personal factors and adverse effects or drug interactions. Patients will ultimately have to decide themselves whether they wish to discontinue drug treatment.

1176 OUTCOME PREDICTION IN PATIENTS WITH BLADDER CANCER AFTER RADICAL CYSTECTOMY USING ARTIFICIAL NEURAL NETWORKS

1176 OUTCOME PREDICTION IN PATIENTS WITH BLADDER CANCER AFTER RADICAL CYSTECTOMY USING ARTIFICIAL NEURAL NETWORKS

Functional Neuroimaging Approaches to the Changing Borders of Consciousness

The bedside diagnosis of vegetative and minimally conscious patients is extremely challenging, and prediction of individual long-term outcome remains difficult. State-of the art neuroimaging methods could help disentangle complex cases and offer new prognostic criteria. These methods can be divided into to three categories: First, new anatomical MRI neuroimaging methods, like diffusion tensor imaging (DTI) or spectroscopy, and passive functional imaging methods (looking at the brain’s activation induced by external stimuli), could provide new diagnostic and prognostic markers. Second, neuroimaging methods based on active collaboration from the patient could help to detect clinically unnoticed signs of consciousness. Third, developments in brain-computer interfaces based on EEG, functional MRI, or EMG offer communication possibilities in brain-damaged patients who can neither verbally nor nonverbally express their thoughts or wishes. These new approaches raise important issues not only from a clinical and ethical perspective (i.e., patients’ diagnosis, prognosis and management) but also from a neuroscientific standpoint, as they enrich our current understanding of the emergence and function of the conscious human mind.

Clinical Algorithm for Improved Prediction of Ambulation and Patient Stratification after Incomplete Spinal Cord Injury

The extent of ambulatory recovery after motor incomplete spinal cord injury (miSCI) differs considerably amongst affected persons. This makes individual outcome prediction difficult and leads to increased within-group variation in clinical trials. The aims of this study on subjects with miSCI were: (1) to rank the strongest single predictors and predictor combinations of later walking capacity; (2) to develop a reliable algorithm for clinical prediction; and (3) to identify subgroups with only limited recovery of walking function. Correlation and logistic regression analyses were performed on a dataset of 90 subjects with tetra- or paraparesis, recruited in a prospective European multicenter study. Eleven measures obtained in the subacute injury period, including clinical examination, tibial somatosensory evoked potentials (tSSEP), and demographic factors, were related to ambulatory outcome (WISCI II, 6minWT) 6 months after injury. The lower extremity motor score (LEMS) alone and in combination was identified as most predictive for later walking capacity in miSCI. Ambulatory outcome of subjects with tetraparesis was correctly predicted for 92% (WISCI II) or 100% (6minWT) of the cases when LEMS was combined with either tSSEP or the ASIA Impairment Scale, respectively. For individuals with paraparesis, prediction was less distinct, mainly due to low prediction rates for individuals with poor walking outcome. A clinical algorithm was generated that allowed for the identification of a subgroup composed of individuals with tetraparesis and poor ambulatory recovery. These data provide evidence that a combination of predictors enables a reliable prediction of walking function and early patient stratification for clinical trials in miSCI.

National Institutes of Health State-of-the-Science Conference Statement: Family History and Improving Health

National Institutes of Health State-of-the-Science Conference Statement: Family History and Improving Health

Individualized Survival and Treatment Response Predictions in Breast Cancer Patients: Involvements of Phospho-EGFR and Phospho-Her2/neu Proteins.

Our robust prediction system for individual breast cancer patients combines three well-known machine-learning classifiers to provide stable and accurate clinical outcome prediction (N=269). The average performance of the selected classifiers is used as the evaluation criterion in breast cancer outcome predictions. A profile (incorporating histology, lymph node status, tumor grade, tumor stage, ER, PR, Her2/neu, patient's age and smoking status) generated over 95% accuracy in individualized disease-free survival and treatment response predictions. Furthermore, our analysis demonstrated that the measurement of phospho-EGFR and phospho-Her2/neu is more powerful in breast cancer survival prediction than that of total EGFR and total Her2/neu (p < 0.05). The incorporation of hormone receptor status, Her2/neu, patient's age and smoking status into the traditional pathologic markers creates a powerful standard to perform individualized survival and treatment outcome predictions for breast cancer patients.

Multiple Gene Expression Classifiers from Different Array Platforms Predict Poor Prognosis of Colorectal Cancer

This study aimed to develop gene classifiers to predict colorectal cancer recurrence. We investigated whether gene classifiers derived from two tumor series using different array platforms could be independently validated by application to the alternate series of patients. Colorectal tumors from New Zealand (n = 149) and Germany (n = 55) patients had a minimum follow-up of 5 years. RNA was profiled using oligonucleotide printed microarrays (New Zealand samples) and Affymetrix arrays (German samples). Classifiers based on clinical data, gene expression data, and a combination of the two were produced and used to predict recurrence. The use of gene expression information was found to improve the predictive ability in both data sets. The New Zealand and German gene classifiers were cross-validated on the German and New Zealand data sets, respectively, to validate their predictive power. Survival analyses were done to evaluate the ability of the classifiers to predict patient survival. The prediction rates for the New Zealand and German gene-based classifiers were 77% and 84%, respectively. Despite significant differences in study design and technologies used, both classifiers retained prognostic power when applied to the alternate series of patients. Survival analyses showed that both classifiers gave a better stratification of patients than the traditional clinical staging. One classifier contained genes associated with cancer progression, whereas the other had a large immune response gene cluster concordant with the role of a host immune response in modulating colorectal cancer outcome. The successful reciprocal validation of gene-based classifiers on different patient cohorts and technology platforms supports the power of microarray technology for individualized outcome prediction of colorectal cancer patients. Furthermore, many of the genes identified have known biological functions congruent with the predicted outcomes.

Prediction of Individual Mandibular Changes Induced by Functional Jaw Orthopedics Followed by Fixed Appliances in Class II Patients

To identify pretreatment cephalometric variables for the prediction of individual mandibular outcomes of functional jaw orthopedics (FJO) followed by fixed appliances in Class II patients treated at the peak in mandibular growth. The study was performed on 51 subjects (24 females, 27 males) with Class II malocclusion. First-phase therapy was accomplished with a twin block in 16 subjects, a stainless steel crown Herbst in 15 subjects, and an acrylic splint Herbst in 20 subjects. Lateral cephalograms were available at the start of treatment with FJO and at the completion of fixed appliance therapy. All subjects received FJO at the peak in mandibular growth (CS 3 at T1). Individual responsiveness to Class II treatment including FJO was defined on the basis of the T2-T1 increment in total mandibular length (Co-Gn) when compared with untreated Class II subjects. Discriminant analysis identified a single predictive parameter (Co-Go-Me degrees) with a classification power of 80%. Pretreatment vertical and sagittal parameters were not able to improve the prediction based upon the mandibular angle. A Class II patient at the peak in skeletal maturation (CS 3) with a pretreatment Co-Go-Me degrees smaller than 125.5 degrees is expected to respond favorably to treatment including FJO. A Class II patient at CS 3 with a pretreatment value for Co-Go-Me degrees greater than 125.5 degrees is expected to respond poorly to treatment including FJO.

Neuro-Fuzzy Modeling: An Accurate and Interpretable Method for Predicting Bladder Cancer Progression

New methods are required to improve the prediction of cancer progression as traditional statistical tests have limited accuracy. Accurate predictions would allow physicians to offer specific treatment according to individual patient risk. While predictive improvements are obtained using ANN, the hidden nature of these networks prevents insight and has hindered their widespread implementation. NFM is an alternate form of artificial intelligence using fuzzy logic (which is a multivalued logic which provides reasoning under uncertainty). By defuzzification the NFM rule base becomes transparent to overcome the black box nature of ANN. Combinations of clinicopathological (tumor stage and grade, patient age, gender, and smoking status) and molecular (immunohistochemical expression of p53 and methylation status of 11 loci) data from 117 patients were used to develop and compare predictive models of tumor progression using NFM, ANN and LR. NFM (88% to 100% sensitivity, 97% to 100% specificity and 94% to 100% accuracy) predicted the presence and timing of cancer progression more accurately than ANN (81% to 87%, 95% to 100% and 89% to 90%, p = 0.002) and LR 3%, 61% to 72% and 47% to 53%, p = 0.00005). NFM was able to interrogate the clinicopathological and molecular data, and select the most important parameters (age, grade, stage, smoking, methylation) for progression prediction. Intelligent systems and molecular biomarkers improved the accuracy of cancer progression predictions. NFM appeared superior to ANN in terms of accuracy, sensitivity, specificity and transparency. The use of NFM in routine clinical practice warrants further validation.

Pro-adrenomedullin to predict severity and outcome in community-acquired pneumonia [ISRCTN04176397

IntroductionPro-adrenomedullin (proADM) is helpful for individual risk assessment and outcome prediction in sepsis. A major cause of sepsis is community-acquired pneumonia (CAP). The aim of this study was to investigate the value of proADM levels for severity assessment and outcome prediction in CAP.MethodsData from 302 patients admitted to the emergency department with CAP were included in a prospective observational study. Procalcitonin, C-reactive protein levels, leukocyte count, clinical variables and the pneumonia severity index (PSI) were measured. ProADM levels were measured with a new sandwich immunoassay for mid regional ProADM (MR-proADM, Brahms AG, Hennigsdorf/Berlin, Germany).ResultsProADM levels, in contrast to C-reactive protein and leukocyte count, increased with increasing severity of CAP, classified according to the PSI score (ANOVA, p < 0.001). In patients who died during follow-up, proADM levels on admission were significantly higher compared to levels in survivors (2.1 (1.5 to 3.0) versus 1.0 (0.6 to 1.6) nmol/l, p < 0.001). In a receiver operating characteristic (ROC) analysis for survival, the area under the ROC curve (AUC) for proADM was 0.76 (95% confidence interval (CI) 0.71–0.81), which was significantly higher compared to procalcitonin (p = 0.004), C-reactive protein (p < 0.001) and total leukocyte count (p = 0.001) and similar to the AUC of the PSI (0.73, p = 0.54). A clinical model including the PSI and proADM increased the prognostic accuracy to predict failure compared to a model relying on the PSI alone (AUC, 0.77 (0.70 to 0.84), p = 0.03).ConclusionProADM, as a novel biomarker, is a useful tool for the risk stratification of patients with CAP.

Early prediction of individual outcome following cardiopulmonary resuscitation: systematic review

Following resuscitation from cardiorespiratory arrest 80% of patients are comatose. Of these patients, 20% will survive and regain consciousness. Is it possible to predict an individual’s long term outcome at...

Gene expression-based, individualized outcome prediction for surgically treated lung cancer patients.

Individualized outcome prediction classifiers were successfully constructed through expression profiling of a total of 8644 genes in 50 non-small-cell lung cancer (NSCLC) cases, which had been consecutively operated on within a defined short period of time and followed up for more than 5 years. The resultant classifier of NSCLCs yielded 82% accuracy for forecasting survival or death 5 years after surgery of a given patient. In addition, since two major histologic classes may differ in terms of outcome-related expression signatures, histologic-type-specific outcome classifiers were also constructed. The resultant highly predictive classifiers, designed specifically for nonsquamous cell carcinomas, showed a prediction accuracy of more than 90% independent of disease stage. In addition to the presence of heterogeneities in adenocarcinomas, our unsupervised hierarchical clustering analysis revealed for the first time the existence of clinicopathologically relevant subclasses of squamous cell carcinomas with marked differences in their invasive growth and prognosis. This finding clearly suggests that NSCLCs comprise distinct subclasses with considerable heterogeneities even within one histologic type. Overall, these findings should advance not only our understanding of the biology of lung cancer but also our ability to individualize postoperative therapies based on the predicted outcome.

Protein S-100b for individual prediction of functional outcome in spinal epidural empyema.

A study correlating protein S-100b serum levels with postoperative functional outcome in patients with spinal cord compression resulting from epidural empyema. To evaluate the potential value of protein S-100b for prediction of individual functional outcome in medullary lesions resulting from spinal epidural empyema. Despite modern medical advances, a reliable individual prediction of functional outcome in case of spinal epidural empyema is still not possible. Forming two outcome groups, clinical outcome following surgery was considered to be favorable in case of neurologic improvement with preservation or retrieval of walking ability, whereas nonimprovement without restoration of gait function was regarded to be unfavorable. Venous blood samples for protein S-100b were taken from all patients immediately after admission and regularly after operative decompression. Initial levels of S-100b were correlated with preoperative degree and duration of paresis, and the individual time course of S-100b measurements was correlated with clinical outcome. The initial level of protein S-100b is not correlated with preoperative degree and duration of paresis or with functional outcome. The individual time course of S-100b measurements, however, is different in both outcome groups. Levels of protein S-100b that were either always normal or that were initially increased but normalized rapidly within 3 days were invariably associated with retrieval of gait function, whereas none of those patients recovered in whom increased levels of S-100b persisted beyond the third postoperative day (P < 0.003). Protein S-100b might be a promising serum marker with prognostic significance in the event of spinal cord compression resulting from epidural empyema.