WO2018223001A1 - Predictive factors for timing of wound closure - Google Patents

Abstract

The present invention relates to methods for treating an open wound. The levels of certain cytokines in serum and wound effluent from a human subject with an open wound are compared to cytokine levels in persons whose wounds were closed successfully or failed to close. The wound is surgically closed if the cytokines levels and wound history of the human subject are favorable. Alternatively, the wound is left open and additional biological samples are analyzed at a later date.

Description

PREDICTIVE FACTORS FOR TIMING OF WOUND CLOSURE

Statement Regarding Federally Sponsored Research or Development

[0001] This invention was made with government support under HT9404-13-1 and H U0001-15-2- 0001 awarded by The Department of Defense. The government has certain rights in the invention.

Background of the Invention

Field of the Invention

[0002] The present invention relates to methods for treating an open wound. The levels of certain cytokines in serum and wound effluent from a human subject with an open wound are compared to cytokine levels in persons whose wounds were closed successfully or failed to close. The wound is surgically closed if the cytokines levels and wound history of the human subject are favorable. Alternatively, the wound is left open and additional biological samples are analyzed at a later date.

Background of the Invention

[0003] The time of closure of combat-related extremity wounds is planned based on the extent of injury and an assessment of local signs of healing such as the absence of tissue necrosis and resolution of infection after multiple surgical debridements (SD). Approximately 80% of these wounds are successfully closed without subsequent complications such as wound failure. Successful wound healing is associated with a distinct local and systemic cytokine response. The present disclosure investigates if the time when a wound should be closed could be estimated based on the cytokine response, traumatic wound characteristics, and occurrence of critical colonization (CC) associated with these injuries.

Introduction

[0004] Extremity wounds are the most common occurring injuries sustained in military casualties, particularly during recent conflicts. These extremity wounds may vary in severity from soft-tissue injuries to comminuted open fractures and traumatic amputations. Explosive blasts are the mechanism of injury in approximately 70% of patients. Considering the extensive amount of tissue damage, initial and ongoing microbiological contamination and further colonization by possible infective agents, regimented treatment of these wounds involves serial surgical debridements every 24 to 48 hours and negative pressure wound therapy, followed by eventual delayed wound closure (DWC) (Connolly M, et al. Changing paradigms in lower extremity reconstruction in war-related injuries. Mil Med Res. 2016;3:9). Wounds are surgically closed by various techniques such as primary closure with approximation of borders by suture, coverage by skin grafts with or without dermal substitutes local, or free myocutaneous or fasciocutaneous flaps. (Sabino J et al. A decade of conflict: flap coverage options and outcomes in traumatic war-related extremity reconstruction. Plast Reconstr Surg. 2015;135(3):895-902; Machen S. Management of traumatic war wounds using vacuum-assisted closure dressings in an austere environment. US Army Med Dep J. 2007:17-23; Geiger S et al., War wounds: lessons learned from Operation Iraqi Freedom. Plast Reconstr Surg. 2008;122(l):146-53).

[0005] The inflammatory response associated with these traumatic wounds is characterized by the production and secretion of several protein biomarkers such as interleukins, chemokines and growth factors both systemically into the blood stream and locally within the wound exudate (Chromy BA et al. Wound outcome in combat injuries is associated with a unique set of protein biomarkers. J Transl Med. 2013;11:281; Hawksworth JS et al. Inflammatory biomarkers in combat wound healing. Ann Surg. 2009;250(6):1002-7). The level of these inflammatory mediators varies during the healing process and may be associated with the extent of injury, wound bioburden, and other complicating factors (Brown TS et al. Inflammatory response is associated with critical colonization in combat wounds. Surg Infect (Larchmt). 2011;12(5):351-7). Additional considerations which may alter the concentration of inflammatory mediators include the development of heterotopic ossification, which is defined by the need of further debridements after attempted definitive wound closure or coverage and wound failure (WF) (Evans KN et al. Inflammatory cytokine and chemokine expression is associated with heterotopic ossification in high-energy penetrating war injuries. J Orthop Trauma. 2012;26(ll):e204-13; Forsberg JA et al. Correlation of procalcitonin and cytokine expression with dehiscence of wartime extremity wounds. J Bone Joint Surg Am. 2008;90(3):580-8). Successful healing is associated with an organized cytokine response, an effective treatment plan based on an adequate number of surgical debridements and efficient use of antibiotics before DWC. The time for DWC is primarily based on the gross appearance of healthy wound tissue, such as proper color, adequate consistency, muscle contractility in response to stimuli, and the capacity to bleed evaluated by examination of wound tissue (Tintle SM et al. Soft tissue coverage of combat wounds. J Surg Orthop Adv. 2010;19(l):29-34; Tintle SM et al, Traumatic and trauma-related amputations: part I: general principles and lower-extremity amputations. J Bone Joint Surg Am. 2010;92(17):2852-68; Tintle SM et al. Traumatic and trauma-related amputations: Part II: Upper extremity and future directions. J Bone Joint Surg Am. 2010;92(18):2934-45; Forsberg JA et al. Do inflammatory markers portend heterotopic ossification and wound failure in combat wounds? Clin Orthop Relat Res.

2014;472(9):2845-54). However, this approach remains subjective and, in approximately 20% of the cases in military and civilian institutions, patients may require reoperation after attempted DWC (Tintle SM et al. Reoperation after combat-related major lower extremity amputations. J Orthop Trauma. 2014;28(4):232-7).

[0006] Bayesian statistics have been used successfully in multivariate analysis studies to assess the relationship of multiple associated variables to estimate the likelihood of patient survival, complications, and treatment outcomes in various fields of medicine (Elster EA et al. Probabilistic (Bayesian) modeling of gene expression in transplant glomerulopathy. J Mol Diagn. 2010;12(5):653- 63; Forsberg JA et al. Estimating survival in patients with operable skeletal metastases: an application of a bayesian belief network. PLoS One. 2011;6(5):el9956; Sohn S et al Detection of clinically important colorectal surgical site infection using Bayesian network. J Surg Res.

2017;209:168-73; Loghmanpour NA et al. A new Bayesian network-based risk stratification model for prediction of short-term and long-term LVAD mortality. ASAIO J. 2015;61(3):313-23; Love A et al. Unifying acute stroke treatment guidelines for a Bayesian belief network. Stud Health Technol Inform. 2013;192:1012; Ashby D. Bayesian statistics in medicine: a 25 year review. Stat Med.

2006;25(21):3589-631).

[0007] In the present study, the timing of DWC for each extremity wound was evaluated by estimating by analyzing possible variations in the level of local and systemic cytokines associated with the inflammatory response and wound healing. When present, these variations in the level of cytokines were hypothesized to accurately predict the timing of successful wound closure. A machine learning approach based on Bayesian statistics was utilized to address this hypothesis. This multivariate analysis approach allows an evaluation of conditional relationships among variables and building of predictive models to estimate the likelihood of an outcome (Forsberg JA et al. Estimating survival in patients with operable skeletal metastases: an application of a bayesian belief network. PLoS One. 2011;6(5):el9956). Using Bayesian statistics, cytokine concentrations and injury characteristics in wounds treated with successful DWC were evaluated to estimate the number of surgical debridements required before these wounds could be closed.

Summary of the Invention

[0008] The time of closure of combat-related extremity wounds is planned based on the extent of injury and wound evolution over multiple surgical debridements. Successful healing is associated with a coordinated local and systemic cytokine response. Therefore, it was investigated whether the time of delayed wound closure (DWC) could be estimated based on the cytokine response and wound characteristics associated with these injuries. 73 patients with 116 wounds were evaluated. Samples of serum and wound effluent were collected during each surgical debridement and tested for 32 cytokines associated with the inflammatory response. Development of critical colonization (CC) was considered as bacterial growth > 10^s CFU per gram of wound tissue or microliter of effluent. A Bayesian Belief Network (BBN) model was trained to estimate the time of DWC. Accuracy and robustness of this model were evaluated by five-fold cross-validation and receiver operator characteristic (ROC) curve. After five-fold cross-validation, a BBN model based on serum levels of IL- 6, IL-7, IL-8, IL-10, and VEGF, effluent levels of IL-lra, IL-5, IL-7, IFN-γ, FGF-basic, MIG, and MIP-loc, injury type, occurrence of multiple wounds, number of days from injury and of previous surgical debridements had successfully estimated the time for immediate DWC (AUC=0.82) and the need for additional debridements (one, AUC=0.6; two, AUC=0.66; three or more, AUC=0.7). The inclusion of additional variables such as ISS, APACHE II, and wound CC status did not improve the performance of the model. The cytokine response in patients with combat-related extremity wounds may aid in predicting time of DWC. These results form the basis of a novel clinical decision support tool to assist in determining a more optimal time to close these wounds.

[0009] The invention encompasses a method of treating an open wound in a human subject comprising: (a) obtaining a first blood, serum or plasma sample from the human subject, (b) determining levels of one or more cytokines selected from interleukin-6 (IL-6), interleukin-7 (IL-7), interleukin-8 (IL-8), interleukin-10 (IL-10) and vascular endothelial growth factor (VEGF) in the first blood, serum or plasma sample, and (c) surgically closing the wound within 24 hours in subjects with cytokine levels about the same as human subjects with successful wound closure. In some embodiments, the method further comprises (d) obtaining a second blood, serum or plasma sample from the human subject at least 24 hours after obtaining the first blood, serum or plasma sample, and (e) determining levels of one or more cytokines selected from IL-6, IL-7, IL-8, IL-10 and VEGF in the second blood, serum or plasma sample.

[0010] The invention encompasses a method of treating an open wound in a human subject comprising: (a) obtaining a first wound effluent sample from the human subject, (b) determining levels of one or more cytokines selected from interleukin-1 receptor agonist (IL-lra), interleukin-5 (IL-5), interleukin-7 (IL-7), basic fibroblast growth factor (FGF-basic), interferon gamma (IFG), monokine-induced gamma interferon (MIG), and macrophage inflammatory protein-1 (MIP-1) in the first wound effluent sample, and (c) surgically closing the wound within 24 hours in subjects with cytokine levels about the same as human subjects with successful wound closure. In some embodiments, the method further comprises (d) obtaining a second wound effluent sample from the human subject at least 24 hours after obtaining the first blood, serum or plasma sample, and (e) determining levels of one or more cytokines selected from IL-lra, IL-5, IL-7, FGF-basic, IFG, MIG, and MIP-1 in the second wound effluent sample.

[0011] In some embodiments, the method further comprises obtaining a first wound effluent sample from the human subject, and determining levels of one or more cytokines selected from IL- lra, IL-5, IL-7, FGF-basic, IFG, MIG, and MIP-1 in the first wound effluent sample. In some embodiments, the method further comprises obtaining a second blood, serum or plasma sample from the human subject at least 24 hours after obtaining the first blood, serum or plasma sample, obtaining a second wound effluent sample from the human subject, determining levels of one or more cytokines selected from IL-6, IL-7, IL-8, IL-10 and VEGF in the second blood, serum or plasma sample, and determining levels of one or more cytokines selected from IL-lra, IL-5, IL-7, FGF-basic, IFG, MIG, and MIP-1 in the second wound effluent sample.

[0012] In some embodiments, the method further comprising debriding the wound. In some embodiments, the method comprises waiting at least two days before surgically closing the wound. In some embodiments, the method further comprises determining the number of open wounds in the human subject, and/or the number of days after the human subject was wounded, and/or the number of times the wound has been debrided. In some embodiments, the method comprises waiting 48 to 72 hours after obtaining the first sample. In some embodiments, the method comprise further debriding the wound. In some embodiments, the wound is surgically closed after determining cytokine levels in the second sample. In some embodiments, the human subject has a trauma wound. In some embodiments, the open wound is an extremity wound.

[0013] The invention also encompasses a method of determining whether to surgically close an open wound comprising: (a) obtaining a blood, serum or plasma sample from the human subject; and (b) measuring levels of one or more cytokines selected from IL-6, IL-7, IL-8, IL-10 and VEGF in the blood, serum or plasma sample; wherein cytokine levels about the same as human subjects with successful wound closure indicate that the wound should be closed within 24 hours.

[0014] The invention encompasses a method of determining whether to surgically close an open wound comprising: (a) obtaining a wound sample from the human subject; and (b) measuring the levels of IL-lra, IL-5, IL-7, FGF-basic, IFG, MIG, and MIP-1 in the wound effluent sample; wherein cytokine levels about the same as human subjects with successful wound closure indicate that the wound should be closed within 24 hours. In some embodiments, the method further comprises (d) obtaining a wound sample from the human subject; and (e) measuring the levels of IL-lra, IL-5, IL-7, FGF-basic, IFG, MIG, and MIP-1 in the wound effluent sample. Brief Description of the Drawings

[0001] Figure 1. Wound closing outcome and cytokine response in patients with combat related extremity wounds A - Kaplan Meier plot of patients with open extremity wounds which healed uneventfully (H) and those which fail to heal (WF) considering the occurrence of delayed wound closure as the event of interest. This plot shows the statistical significant difference confirmed by the Mantel-Cox test (p<0.05). B - Pictures of two different patients with extremity wounds on day of delayed wound closure (DWC) showing apparently similar signs of healing. The right lower extremity (RLE) wound healed uneventfully and the left lower extremity (LLE) failed to heal. C and D- Level of cytokines in the serum (C) and in the wound effluent (D) of patients at admission and on day of DWC. The level of cytokines on day of DWC was separated according to successful healing (H) or failure to heal (WF). Cytokine level was evaluated using the Kruskal-Wallis test, followed by Dunn's Multiple Comparison after confirming data distribution with a Shapiro-Wilk normality test and p- values were adjusted using the false discovery rate (FDR) controlled with a Benjamini-Yekutieli adjustment. * Represent a statistical significant difference (p<0.05) as indicated.

[0002] Figure 2. Biomarker distributions for 32 analytes compared between patients with healed and dehisced wounds. (A) Serum biomarkers from all time points. (B) Serum biomarkers from the final time point. (C) Effluent biomarkers from all time points. (D) Effluent biomarkers from the final time point. The ends of each box mark upper and lower quartiles. The median is marked by a horizontal line. The whiskers extend to the highest and lowest observations.

[0003] Figure 3. Injury type and the occurrence of multiple extremity wounds. A - Kaplan Meier plot of patients with open fractures, soft tissue injures and amputations considering the occurrence of delayed wound closure as the event of interest. This plot shows the statistical significant difference confirmed by the Mantel-Cox test (p<0.05). B - Bar graph of number of wounds distributed by injury type considering occurrence of single or multiple extremity wounds. The association of these variables is confirmed with a Fisher test and statistical significant difference (p<0.05) is shown as indicated. C - Kaplan Meier plot of patients with single and multiple extremity wounds considering the occurrence of delayed wound closure as the event of interest. This plot shows the statistical significant difference confirmed by the Mantel-Cox test (p<0.05). D - Bar graph of number of single and multiple wounds considering successful healing (H) or failure to heal (WF). The association of these variables is confirmed with a Fisher test and statistical significant difference (p<0.05) is shown as indicated. [0004] Figure 4. Baysian Belief Network model estimating the number of debridements for DWC. Diagrams represent the distribution of each predictor variable and the connections represent the degree of association of predictors with the outcome, highlighted in the center of the model. Based on the association with the time of wound closure in a Naive Bayes classifier, cytokines were selected to train a Bayesian Belief Network (BBN) model and estimate the time of DWC in 90 wounds with successful closure. This BBN model was developed using machine-learning algorithms and controlled for the number of previous debridement surgeries and the number of days from injury. These algorithms are based on a scoring formula which balances goodness-of-fit and robustness using a parsimony metric to reduce the risk of overfitting the model to the training set.

[0005] Figure 5. Receiver operating characteristic (ROC) curve of the number of additional surgical debridements estimated to successfully close combat-related extremity wounds (BBN model described in Figure 3). DWC represent immediate closure without additonal surgical debridements. The combined area under the curve (AUC) was estimated for immediate DWC (0.82), one (0.6), two (0.66), and three or more (0.7) additional surgical debridements by five-fold cross-validation.

Detailed Description of the Invention

[0006] The present invention relates to methods of determining if a su bject has an increased risk of developing delayed wound closure (DWC) prior to the onset of any detectable symptoms thereof. The methods comprise analyzing at least one sample from the subject to determine a value of the subject's risk profile and comparing the value of the subject's risk profile with the value of a normal risk profile. A change in the value of the subject's risk profile, over or under normal values is indicative that the subject has an increased risk of having or developing symptoms associated with DWC prior to the onset of any detectable symptoms thereof.

[0007] The present invention also relates to methods of detecting elevated levels of a specific collection of analytes in one or more samples obtained from a subject. In one embodiment, the collection of analytes comprises serum levels of interleukin-6 (I L-6), interleukin-7 (I L-7), interleukin-8 (I L-8), interleukin-10 (I L-10) and vascular endothelial growth factor (VEG F). In another embodiment, the collection of analytes comprises wound effluent levels of interleukin-lra (IL-lra), interleukin-5 (I L-5), interleukin-7 (IL-7), basic fibroblast growth factor (FGF-basic), interferon gamma (I FG), monokine-induced gamma interferon (M IG), macrophage inflammatory protein-1 (M I P-1).

[0008] As used herein, the term "subject" or "test subject" indicates a mammal, in particular a human or non-human primate. The test subject is in need of an assessment of susceptibility of DWC. For example, the test subject may have no symptoms that DWC may occur. [0009] Use of the term "about" when referring to a number or a numerical range means that the number or numerical range referred to is an approximation within experimental variability (or within statistical experimental error), and thus the number or numerical range may vary from, for example, between 1% and 15% of the stated number or numerical range.

[0010] The term delayed wound closure, or DWC, is used herein to mean closure of a wound in a delayed fashion, or by tertiary intention. Wound closure is delayed due to the complexity of the wound, which necessitates a more complex treatment regimen to include surveillance for infection, monitoring of wound bed formation, and serial sharp surgical debridements. DWC includes wounds closed using delayed primary closure and wounds closed using va rious grafting and complex closure techniques.

[0011] As used herein, the term means "increased risk" is used to mean that the test subject has an increased chance of developing DWC compared to a normal individual. The increased risk may be relative or absolute and may be expressed qualitatively or quantitatively. For example, an increased risk may be expressed as simply determining the subject's risk profile and placing the patient in an "increased risk" category, based upon previous population studies. Alternatively, a numerical expression of the subject's increased risk may be determined based upon the risk profile. As used herein, examples of expressions of an increased risk include but are not limited to, odds, probability, odds ratio, p-values, attributable risk, biomarker index score, relative frequency, positive predictive value, negative predictive value, and relative risk.

[0012] For example, the correlation between a subject's risk profile and the likelihood of developing DWC may be measured by an odds ratio (OR) and by the relative risk (RR). If P(R⁺) is the probability of developing DWC for individuals with the risk profile (R) and P(R^") is the probability of developing DWC for individuals without the risk profile, then the relative risk is the ratio of the two probabilities: RR=P(R⁺)/P(R-).

[0013] In case-control studies, however, direct measures of the relative risk often cannot be obtained because of sampling design. The odds ratio allows for an approximation of the relative risk for low-incidence diseases and can be calculated: OR=(F⁺/( l-F⁺))/(F^"/(l-F^")), where F⁺ is the frequency of a risk profile in cases studies and F^" is the frequency of risk profile in controls. F⁺ and F^" can be calculated using the risk profile frequencies of the study.

[0014] The attributable risk (AR) can a lso be used to express an increased risk. The AR describes the proportion of individuals in a population exhibiting DWC to a specific member of the risk profile. AR may also be important in quantifying the role of individual components (specific member) in condition etiology and in terms of the public health impact of the individual risk factor. The public health relevance of the AR measurement lies in estimating the proportion of cases of DWC in a population of wounded subjects that could be prevented if the profile or individual factor were absent. AR may be determined as follows: AR=P_e(RR-1)/(PE(RR-1)+1), where AR is the risk attributable to a profile or individual factor of the profile, and PE is the frequency of exposure to a profile or individual component of the profile within the population at large. RR is the relative risk, which can be approximated with the odds ratio when the profile or individual factor of the profile under study has a relatively low incidence in the general population.

[0015] In one embodiment, the increased risk of a patient can be determined from p-values that are derived from association studies. Specifically, associations with specific profiles can be performed using regression analysis by regressing the risk profile with the presence or absence of DWC. In addition, the regression may or may not be corrected or adjusted for one or more factors. The factors for which the analyses may be adjusted include, but are not limited to age, sex, weight, ethnicity, type of wound, number of wounds, number of days from injury, number of previous surgical debridements, geographic location, fasting state, state of pregnancy or post-pregnancy, menstrual cycle, general health of the subject, alcohol or drug consumption, caffeine or nicotine intake and circadian rhythms, to name a few.

[0016] Increased risk can also be determined from p-values that are derived using logistic regression. Binomial (or binary) logistic regression is a form of regression which is used when the dependent is a dichotomy and the independents are of any type. Logistic regression can be used to predict a dependent variable on the basis of continuous or categorical or both (continuous and categorical) independents and to determine the percent of variance in the dependent variable explained by the independents; to rank the relative importance of independents; to assess interaction effects; and to understand the impact of covariate control variables. Logistic regression applies maximum likelihood estimation after transforming the dependent into a "logit" variable (the natural log of the odds of the dependent occurring or not). In this way, logistic regression estimates the probability of a certain event occurring. These analyses may be conducted with virtually any statistics program, such as not limited to SAS, R package available through CRAN repository.

[0017] SAS ("statistical analysis software") is a general purpose package (similar to Stata and SPSS) created by Jim Goodnight and N.C. State University colleagues. Ready-to-use procedures handle a wide range of statistical analyses, including but not limited to, analysis of variance, regression, categorical data analysis, multivariate analysis, survival analysis, psychometric analysis, cluster analysis, and nonparametric analysis. R package is free, general purpose package that complies with and runs on a variety of UNIX platforms.

[0018] Accordingly, select embodiments of the present invention comprise the use of a computer comprising a processor and the computer is configured or programmed to generate one or more risk profiles and/or to determine statistical risk. The methods may also comprise displaying the one or risk profiles on a screen that is communicatively connected to the computer. In another embodiment, two different computers can be used: one computer configured or programmed to generate one or more risk profiles and a second computer configured or programmed to determine statistical risk. Each of these separate computers can be communicatively linked to its own display or to the same display.

[0019] As used herein, the phrase "risk profile" means the combination of a subject's risk factors analyzed or observed. The terms "factor" and/or "component" are used to mean the individual constituents that are assessed when generating the profile. The risk profile is a collection of measurements, such as but not limited to a quantity or concentration, for individual factors taken from a test sample of the subject. Examples of test samples or sources of components for the risk profile include, but are not limited to, biological fluids, which can be tested by the methods of the present invention described herein, and include but are not limited to whole blood, such as but not limited to peripheral blood, serum, plasma, cerebrospinal fluid, wound effluent, urine, amniotic fluid, lymph fluids, various external secretions of the respiratory, intestinal and genitourinary tracts, tears, saliva, white blood cells, myelomas and the like.

[0020] The risk profile can include a "biological effector" aspect and/or a non-biological effector aspect. As used herein, the term "biological effector" is used to mean a molecule, such as but not limited to, a protein, peptide, a carbohydrate, a fatty acid, a nucleic acid, a glycoprotein, a proteoglycan, etc. that can be assayed. Specific examples of biological effectors can include, cytokines, growth factors, antibodies, hormones, cell surface receptors, cell surface proteins, carbohydrates, etc. More specific examples of biological effectors include interleukins (ILs) such as IL-la, IL-Ιβ, IL-1 receptor antagonist (IL-1RA), IL-2, IL-2 receptor (IL-2R), IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12, IL-13, IL-15, IL-17, as well as growth factors such as tumor necrosis factor alpha (TNFa), granulocyte colony stimulating factor (G-CSF), granulocyte macrophage colony stimulating factor (GM-CSF), interferon alpha (INF-a), interferon gamma (IFN-y), epithelial growth factor (EGF), basic endothelial growth factor (bEGF), hepatocyte growth factor (HGF), vascular endothelial growth factor (VEGF), and chemokines such as monocyte chemoattractant protein-1 (CCL2/MCP-1), macrophage inflammatory protein-1 alpha (CCL3/MIP-la), macrophage inflammatory protein-1 beta (CCI-4/Μ Ι Ρ-Ιβ), CCL5/RANTES, CCLll/eotaxin, monokine induced by gamma interferon (CXCL9/MIG) and interferon gamma-induced protein-10 (CXCL10/I P10).

[0021] Techniques to assay levels of individual components of the risk profile from test samples are well known to the skilled technician, and the invention is not limited by the means by which the components are assessed. In one embodiment, levels of the individual factors in the serum or wound effluent of the risk profile are assessed using mass spectrometry in conjunction with ultra- performance liquid chromatography (U PLC), high-performance liqu id chromatography (H PLC), gas chromatography (GC), gas chromatography/mass spectroscopy (GC/MS), and UPLC to name a few. Other methods of assessing levels of some of the individual components include biological methods, such as but not limited to ELISA assays, Western Blot and multiplexed immunoassays etc. Other techniques may include using quantitative arrays, PCR, Northern Blot analysis. To determine levels of components or factors, it is not necessary that an entire component, e.g., a full length protein or an entire RNA transcript, be present or fully sequenced. In other words, determining levels of, for example, a fragment of protein being analyzed may be sufficient to conclude or assess that an individual component of the risk profile being analyzed is increased or decreased. Similarly, if, for example, arrays or blots are used to determine component levels, the presence/absence/strength of a detectable signal may be sufficient to assess levels of components.

[0022] As used herein, the term non-biological effector is a component that is generally considered not to be a specific molecule. Although not a specific molecule, a non-biological effector may nonetheless still be quantifiable, either through routine measurements or through measurements that stratify the data being assessed. For example, number or concentrate of red blood cells, white blood cells, platelets, coagulation time, blood oxygen content, etc. would be a non-biological effector component of the risk profile. All of these components are measureable or quantifiable using routine methods and equipment. Other non-biological components include data that may not be readily or routinely quantifiable or that may require a practitioner's judgment or opinion. For example, wound severity may be a component of the risk profile. While there may be published guidance on classifying wound severity, stratifying wound severity and, for example, assigning a numerical value to the severity, still involves observation and, to a certain extent, judgment or opinion. In some instances the quantity or measurement assigned to a non-biological effector could be binary, e.g., "0" if absent or "1" if present. In other instances, the non-biological effector aspect of the risk profile may involve qualitative components that cannot or should not be quantified.

[0023] In one embodiment, the mechanism of injury is included in the risk profile. As used herein, the phrase "mechanism of injury" means the manner in which the subject received an injury. For example, the mechanism of injury may be described as a gunshot wound, a vehicle accident, laceration, etc. In another embodiment, data regarding injury type is included in the risk profile. In another embodiment, data on the occurrence of multiple wounds is included in the risk profile. In another embodiment, data on the number of days from injury is included in the risk profile. In another embodiment, data on the number of surgical debridements (SDs).

[0024] Other examples of individual components of the risk profile include but are not limited to ISS score of head, ISS score of chest (thorax) and critical colonization. "ISS score" (injury severity score) is well-known in the art and is used routinely in clinics to assess severity of wounds or injuries.

[0025] As used herein, the term critical colonization (or "CC") is a measure of CFU that the subject has in serum and/or tissue for at least one wound when initially examined by the attending physician. For example, if a subject has CFU of lxlO⁵ per ml of serum, or if at least one wound has CFU of 1x10^s per mg of tissue, the subject is said to be "positive" for CC. If the total serum CFU or no single would has CFU of at least lxlO⁵ the subject is said to be "negative" for CC.

[0026] To determine which of the biological effector or non-biological effector components may be critical in the subjects' risk profiles, routine statistical methods can be employed. For example, rflmpute from the randomForest R package can be used to impute missing data. Up-sampling and predictor rank transformations can be performed on the data set only for variable selection to accommodate class imbalance and non-normality in the data.

[0027] For variable selection, the constraint-based algorithms /asi./amfa, iamb and gs and the constraint-based local discovery learning algorithms mmpc and si.hiton.pc from the "bnlearn" R package can be used to search the input dataset for nodes of Bayesian networks. The nodes can be chosen as the reduced variable sets. Before running the data through variable selection and binary classification algorithms, the variables may or may not randomly re-ordered. The data can be run through the variable selection and binary classification algorithms more than once, for example, 10, 20, 30, 40, 50 or even more times.

[0028] For binary classification and model selection, each variable set can be pulled from the raw data and run in sundry binary classification algorithms using the train function from the R caret package: linear discriminant analysis {Ida), classification and regression trees {cart), k-nearest neighbors {knn), support vector machine {svm), logistic regression {glm), random forest {rf), generalized linear models {glmnet) and naive Bayes {nb). The best variable set and binary classification algorithm combination that first produces the highest kappa and then the highest sensitivity with reasonable specificity can then be chosen. [0029] The resultant models are then examined using accuracy, no information rate, positive predictive value and negative predictive value. Model performance can be further assessed using the plot. roc command to compute the Receiver Operator Characteristic Curves (ROC) and area under curve (AUC). The dca R command from the Memorial Sloan Kettering Cancer Center website, www.mskcc.org, can be used to compute the Decision Curve Analysis (DCA).

[0030] Finally, for univariate analysis, a Wilcoxon rank-sum test can be used to identify which biomarkers from specific patient groups are were associated with a specific indication.

[0031] The assessment of the levels of the individual components of the risk profile can be expressed as absolute or relative values and may or may not be expressed in relation to another component, a standard an internal standard or another molecule of compound known to be in the sample. If the levels are assessed as relative to a standard or internal standard, the standard may be added to the test sample prior to, during or after sample processing.

[0032] To assess levels of the individual components of the risk profile, a sample may be taken from the subject. The sample may or may not processed prior assaying levels of the components of the risk profile. For example, whole blood may be taken from an individual and the blood sample may be processed, e.g., centrifuged, to isolate plasma or serum from the blood. The sample may or may not be stored, e.g., frozen, prior to processing or analysis.

[0033] In one embodiment, the individua l levels of each of the risk factors are higher or lower than those compared to normal levels. In another embodiment, one, two, three, four, five, six or seven of the levels of each of the factor are higher or lower than normal levels while others, if any, are lower than or the same as normal levels.

[0034] The levels of depletion of the factors or components compared to normal levels can vary. In one embodiment, the levels of any one or more of the factors or components is at least 1.05, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 higher than normal levels (where, for sake of clarity, a marker a level of "1" would indicate that the component is at the same level in both the subject and normal samples). For the purposes of the present invention, the number of "times" the levels of a factor are higher over normal can be a relative or absolute number of times. In the alternative, the levels of the factors or components may be normalized to a standard and these normalized levels can then be compared to one another to determine if a factor or component is lower, higher or about the same. [0035] For the purposes of the present invention the risk profile comprises at least one, two, three, four, five, six or seven of the factors or components for the prediction of DWC. If one factor or component of the biological effector aspect of the risk profile is used in generating the risk profile for the prediction of DWC, then any one of the listed factors or components can be used to generate the profile. If two factors or components of the biological effector aspect of the risk profile are used in generating the risk profile for the prediction of DWC, any combination of the two listed above can be used. If three factors or components of the biological effector aspect of the risk profile are used in generating the risk profile for the prediction of DWC, any combination of three of the factors or components listed above can be used. If four factors or components of the biological effector aspect of the risk profile are used in generating the risk profile for the prediction of DWC, any combination of four of the factors or components listed above can be used. If five factors or components of the biological effector aspect of the risk profile are used in generating the risk profile for the prediction of DWC, any combination of five of the factors or components listed above can be used. If six factors or components of the biological effector aspect of the risk profile are used in generating the risk profile for the prediction of DWC, any combination of six of the factors or components listed above can be used. Of course all members or components of the biological effector aspect of the risk profile can be used in generating the risk profile for the prediction of DWC.

[0036] The subject's risk profile is compared to the profile that is deemed to be a normal risk profile. To establish the risk profile of a normal individual, an individual or group of individuals may be first assessed to ensure they have no signs, symptoms or diagnostic indicators of DWC. Once established, the risk profile of the individual or group of individuals can then be determined to establish a "normal risk profile." In one embodiment, a normal risk profile can be ascertained from the same subject when the subject is deemed as healthy with no signs, symptoms or diagnostic indicators of DWC. In one embodiment, a risk profile from a "normal subject," e.g., a "normal risk profile," is a subject with a wound but did not exhibit or display DWC. In one specific embodiment, the normal subject has a chest wound, head wound or as extremity (arm, hand, finger(s), leg, foot, toe(s)) wound but did not exhibit DWC.

[0037] In one embodiment, a "normal" risk profile is assessed in the same subject from whom the sample is taken prior to the onset of any signs, symptoms or diagnostic indicators that they may exhibit DWC. That is, the term "normal" with respect to a risk profile can be used to mean the subject's baseline risk profile prior to the onset of any wounds, signs, symptoms or diagnostic indicators of potential DWC. The risk profile can then be reassessed periodically and compared to the subject's baseline risk profile. Thus, the present invention also includes methods of monitoring the progression of DWC in a subject, with the methods comprising determining the subject's risk profile at more than one time point. For example, some embodiments of the methods of the present invention will comprise determining the subject's risk profile at two, three, four, five, six, seven, eight, nine, 10 or even more time points over a period of time, such as a week or more, two weeks or more, three weeks or more, four weeks or more, a month or more, two months or more, three months or more, four months or more, five months or more, six months or more, seven months or more, eight months or more, nine months or more, ten months or more, 11 months or more, a year or more or even two years. The methods of monitoring a subject's risk of developing DWC would also include embodiments in which the subject's risk profile is assessed before and/or during and/or after treatment of DWC. In other words, the present invention also includes methods of monitoring the efficacy of treatment of DWC by assessing the subject's risk profile over the course of the treatment and after the treatment. In specific embodiments, the methods of monitoring the efficacy of treatment of DWC comprise determining the subject's risk profile at at least one, two, three, four, five, six, seven, eight, nine or 10 or more different time points prior to the receipt of treatment for DWC and subsequently determining the subject's risk profile at at least one, two, three, four, five, six, seven, eight, nine or 10 or more different time points after beginning of treatment for DWC, and determining the changes, if any, in the risk profile of the subject. The treatment may be any treatment designed to cure, remove or diminish the likelihood of developing DWC.

[0038] In another embodiment, a normal risk profile is assessed in a sample from a different subject or patient (from the subject being analyzed) and this different subject does not have or is not suspected of developing DWC. In still another embodiment, the normal risk profile is assessed in a population of healthy individuals, the constituents of which display no signs, symptoms or diagnostic indicators that they may have or will develop DWC. Thus, the subject's risk profile can be compared to a normal risk profile generated from a single normal sample or a risk profile generated from more than one normal sample.

[0039] Of course, measurements of the individual components, e.g., concentration, ratio, log ratios etc., of the normal risk profile can fall within a range of values, and values that do not fall within this "normal range" are said to be outside the normal range. These measurements may or may not be converted to a value, number, factor or score as compared to measurements in the "normal range." For example, a measurement for a specific factor or component that is below the normal range, may be assigned a value or -1, -2, -3, etc., depending on the scoring system devised.

[0040] In another embodiment, the measurements of the individual components themselves are used in the risk profile, and these levels can be used to provide a "binary" value to each component, e.g., "elevated" or "not elevated." Each of the binary values can be converted to a number, e.g., "1" or "0," respectively.

[0041] In one embodiment, the "risk profile value" can be a single value, number, factor or score given as an overall collective value to the individual components of the profile. For example, if each component is assigned a value, such as above, the component value may simply be the overall score of each individual or categorical value. For example, if five components of the risk profile for predicting DWC are used and three of those components are assigned values of "+2" and two are assigned values of the risk profile in this example would be +8, with a normal value being, for example, "0." In this manner, the risk profile value could be a useful single number or score, the actual value or magnitude of which could be an indication of the actual risk of developing DWC, e.g., the "more positive" the value, the greater the risk of developing DWC.

[0042] In another embodiment the "risk profile value" can be a series of values, numbers, factors or scores given to the individual components of the overall profile. In another embodiment, the "risk profile value" may be a combination of values, numbers, factors or scores given to individual components of the profile as well as values, numbers, factors or scores collectively given to a group of components, such as a biological effector portion. In another example, the risk profile value may comprise or consist of individual values, number or scores for specific component as well as values, numbers or scores for a group of components.

[0043] In another embodiment individual values from the risk profile and/or the mechanism of injury can be used to develop a single score, such as a "combined risk index," which may utilize weighted scores from the individual component values reduced to a diagnostic number value. The combined risk index may also be generated using non-weighted scores from the individual component values. When the "combined risk index" exceeds a specific threshold level, determined by a range of values developed similarly from control (normal) subjects, the individual has a high risk, or higher than normal risk, of developing DWC, whereas maintaining a normal range value of the "combined risk index" would indicate a low or minimal risk of developing DWC. In this embodiment, the threshold value would be or could be set by the combined risk index from one or more normal subjects.

[0044] In another embodiment, the value of the risk profile can be the collection of data from the individual measurements and need not be converted to a scoring system, such that the "risk profile value" is a collection of the individual measurements of the individual components of the profile. [0045] In specific embodiments, a subject is diagnosed of having an increased risk of suffering from DWC if the subject's seven, six, five, four, three, two or even one of the components or factors herein are at abnormal levels.

[0046] If it is determined that a subject has an increased risk of developing DWC, the attending health care provider may subsequently prescribe or institute a treatment program. In this manner, the present invention also provides for methods of treating individuals for DWC. The attending healthcare worker may begin treatment, based on the subject's risk profile, before there are perceivable, noticeable or measurable signs of DWC in the individual.

[0047] Accordingly, the invention provides methods of treating DWC in a subject in need thereof. The treatment methods include obtaining a subject's risk profile as defined herein and prescribing a treatment regimen to the subject if the risk profile indicates that the subject is at risk of developing DWC.

[0048] The methods of treatment also include methods of monitoring the effectiveness of a treatment for DWC. Once a treatment regimen has been established, with or without the use of the methods of the present invention to assist in a diagnosis of a risk of developing DWC, the methods of monitoring a subject's risk profile over time can be used to assess the effectiveness of treatments for DWC. Specifically, the subject's risk profile can be assessed over time, including before, during and after treatments for DWC. The risk profile can be monitored, with, for example, the normalization or decline in the values of the profile over time being indicative that the treatment may be showing efficacy of treatment.

[0049] The present invention also provides kits that can be used in the methods of the present invention. Specifically, the present invention provides kits for assessing the increased risk of developing DWC, with the kits comprising one or more sets of antibodies that are immobilized onto a solid substrate and specifically bind to at least one of the factors or components listed herein. In specific embodiments, the kits comprise at least two, three, four, five, six or seven sets of antibodies immobilized onto a solid substrate, with each set corresponding to a factor.

[0050] The antibodies that are immobilized onto the substrate may or may not be labeled. For example, the antibodies may be labeled, e.g., bound to a labeled protein, in such a manner that binding of the specific protein may displace the label and the presence of the marker in the sample is marked by the absence of a signal. In addition, the antibodies that are immobilized onto the substrate may be directly or indirectly immobilized onto the surface. Methods for immobilizing proteins, including antibodies, are well-known in the art, and such methods may be used to immobilize a target protein, e.g., I L-10, or another antibody onto the surface of the substrate to which the antibody directed to the specific factor can then be specifically bound. In this manner, the antibody directed to the specific biomarker is immobilized onto the surface of the substrate for the purposes of the present invention.

[0051] The kits of the present invention may or may not include containers for collecting samples from the subject and one or more reagents, e.g., purified target biomarker for preparing a calibration curve. The kits may or may not include additional reagents such as wash buffers, labeling reagents and reagents that are used to detect the presence (or absence) of the label.

[0052] All patents and publications cited herein are incorporated by reference to the same extent as if each individual publication was specifically and individually indicated as having been incorporated by reference in its entirety.

Examples

Example 1: Methods

[0053] Study Population : This study evaluated United States combat casualties with extremity wounds, enrolled in an Institutional Review Board-approved study, who were treated at the Walter Reed National Military Medical Center (WRN MMC) between 2007 and 2012. Patients were initially treated at U.S. military medical facilities in the area of combat operations and eventually transferred to WRN M MC. As described previously ( Lisboa FA et al. Nonsteroidal anti-inflammatory drugs may affect cytokine response and benefit healing of combat-related extremity wounds. Surgery.

2017;161(4): 1164-73), patient treatment followed guidelines from the Department of Defense Joint Trauma System. Throughout this study, military casualties were evacuated and treated based on the "golden hour" mandate by the Secretary of Defense Robert M. Gates. (Morrison JJ et al. En-route care capability from point of injury impacts mortality after severe wartime injury. Ann Surg.

2013;257(2):330-4; Kotwal RS et al. The Effect of a Golden Hour Policy on the Morbidity and Mortality of Combat Casualties. JAMA Surg. 2016; 151(l): 15-24) Patients were aeromedically evacuated back to the U.S. and underwent surgical debridements at staged facilities along the continuum of care prior to arrival to the WRN M MC, which occurred within three to six days after injury (average of 4.9 days). All patients signed informed consent forms at the time of their enrollment. Data regarding injury mechanism and demographics was collected from all participating patients. This data included age, body mass index (BMI), tobacco use, treatment time until definitive wound closure in number of days, Injury Severity Score (ISS), Acute Physiology and Chronic Health Evaluation I I (APACH E I I) score, critical colonization (defined as bacterial growth >10⁵CFU per gram of tissue or microliter of wound effluent) of wounds, wound surface area, wound size, systemic cytokine profile, and wound healing outcome. Wound failure (WF) was defined as infection, dehiscence (spontaneous partial or complete disruption after attempted delayed primary closure) or loss of >10% of a skin graft requiring a return to the operating room. If patients did not return to the operating room after 30 days of definitive closure, wounds were regarded as successfully healed (Hawksworth JS et al. Inflammatory biomarkers in combat wound healing. Ann Surg. 2009;

250(6): 1002-7).

[0054] Sample Collection: Sample collection was performed as previously described. (Hawksworth JS et al. Inflammatory biomarkers in combat wound healing. Ann Surg. 2009; 250(6): 1002-7) Briefly, a sample of peripheral venous blood (~8 ml) and a volume of approximately 20 to 30 ml of wound effluent were collected at the operating room before each surgical debridement. Wound effluent samples were collected by negative pressure wound therapy during a period of 12 hours prior to each debridement. Next, blood and wound effluent samples were centrifuged at 2,500 x g at 4°C for 10 minutes. After centrifugation, supernatant aliquots of serum and wound effluent were transferred to Cryo-Loc polypropylene tubes and labeled (Lake Charles Manufacturing, Lake Charles, LA). In the sequence, aliquots were flash-frozen in liquid nitrogen and stored at -80°C for batch analysis. Wound pictures obtained prior to each debridement procedure at the operating room were analyzed using the PictZar^® Pro digital planimetry software (PictZar^®, Elmwood Park, NJ) to calculate wound surface area.

[0055] Cytokine Analysis: On day of testing, frozen aliquots of serum and wound effluent were thaw and filtered using a 0.65μιη filter (Millipore, Billerica, MA). Next, samples were combined in 96 plates and tested with a Human Cytokine 30-plex panel kit supplemented with a custom Human 2- plex panel (Invitrogen; Cat. No LH6003 and LCP0002) for the following 32 cytokines: I L-la, I L-Ιβ, IL- lra, I L-2, I L-2R, I L-3, I L-4, I L-5, I L-6, I L-7, IL-8, I L-10, IL-12, I L-13, IL-15, I L-17, GM-CSF, G-CSF, I N F-ffl, I N F-ffl, TN F-ffl, EGF, bFGF, HG F, VEGF, Eotaxin, MCP-1, ΜΙ Ρ-Ια, ΜΙ Ρ-Ιβ, RANTES, M IG, and I P-10. Plates were analyzed using a multiplex analysis platform (Luminex 100 System, Luminex, Austin, TX) and the software BeadView (Upstate VI.0.4.23259, Millipore, Billerica, MA).

[0056] Quantitative Bacteriology: A wound tissue biopsy sample (~1 cm3) was obtained from the center of each wound during all SD. Next, samples were weighed and placed in sterile 15-m L conical vials at 4°C for culture. As described previously, (Lisboa FA et al. Bilateral lower-extremity amputation wounds are associated with distinct local and systemic cytokine response. Surgery. 2013; 154(2):282-90) tissue samples were transferred to a sterile grinder and diluted 1: 10 (wt/vol) in fastidious broth to reach a homogeneous final concentration of 0.1 gram of tissue per milliliter. In the sequence, all samples were inoculated on sheep's blood agar and MacConkey agar plates in triplicate. Culture plates were incubated at 37°C overnight and on the next day colonies were counted. The total number of CFUs in culture plates per gram of tissue or microliter of wound effluent was calculated. The phenotypic identification of colon ies was completed by using an automated bacterial identification system (Phoenix, Becton Dickinson, Sparks, MD). A bacterial growth > 10⁵CFU per gram of tissue or microliter of wound effluent was considered as critical colonization (CC).

[0057] Statistical Analysis: The Shapiro-Wilk test was used to evaluate data distribution. Categorical variables such as wound closure outcome, wound type (defined as soft tissue injuries (STI), open fractures (OF) and amputations (A)), occurrence of multiple extremity wounds and development of CC were analyzed using Kaplan-Meier estimates. Typically, Kaplan-Meier estimates were used considering mortality as the event of interest. In the present analysis, wound closure was considered as the event of interest. The associations between these categorical variables were further evaluated with Chi-square test. Cytokine level was evaluated using the Kruskal-Wallis test, followed by Dunn's Multiple Comparison after confirming data distribution with a Shapiro-Wilk normality test. The p-values were adjusted using the false discovery rate (FDR) controlled with a Benjamini-Yekutieli adjustment (Benjamini Y and Yekutieli D. The Control of the False Discovery Rate in Multiple Testing under Dependency. The Annals of Statistics. 2001;29(4): 1165-88). A two-step process was used to estimate the time of DWC. First, a Bayes classifier was trained to identify the association of cytokines with the time of DWC in 90 wounds with favorable healing and successful closure. Next, the identified cytokines, the number of previous surgical debridements and the number of days from injury were used to train a Bayesian Belief Network ( BBN) using the data from these same patients with 90 wounds and successful closure. The Naive Bayes classifier and BBN models were developed using commercially available machine-learning software (FasterAnalytics, DecisionQ, Washington, DC) as described previously (Forsberg JA et al. Estimating survival in patients with operable skeletal metastases: an application of a bayesian belief network. PLoS One. 2011;6(5):el9956). In brief, joint probability distributions— that is, how and under what circumstances features relate to one another— are not specified a priori, but instead learned from the training set (Nandra R et al. Can a Bayesian Belief Network Be Used to Estimate 1-year Survival in Patients With Bone Sarcomas? Clin Orthop Relat Res. 2017;475(6) :1681-9; Forsberg JA et al. Can We Estimate Short- and Intermediate- term Survival in Patients Undergoing Surgery for Metastatic Bone Disease? Clin Orthop Relat Res. 2017;475(4): 1252-61). The accuracy and robustness of this model were evaluated by five-fold cross- validation and the calculated area under the receiver operator characteristic curve (AUC). The spearman's correlation coefficient and 95%CI were calculated to evaluate the association of each cytokine in the model with the timing of DWC represented by the number of surgical debridements needed for successful wound closure. Statistical analyses were performed using the RStudio software (Version 0.98.1103, RStudio, Inc.) and GraphPad Prism software (Version 7, GraphPad Software, Inc). Serum analysis and outcomes were considered per patient, as opposed to per wound, unless stated otherwise and a p<0.05, after FDR adjustment, was regarded as statistically significant.

[0058] Univariate statistical analysis was performed as appropriate to select possible additional associated wound characteristics for the BBN model and p<0.05 was considered statistically significant.

Example 2

[0059] A total of 73 combat casualties with 116 wounds were enrolled over the six-year study period. Patients had, on average, at least two surgical debridements after admission to WRNMMC and a mean ISS between 19 and 20. Of these patients, 60 with 90 wounds had successful DWC without the need for further interventions. In these patients with successful healing, DWC occurred at later dates from injury as compared to those that had wound failure (Figure 1A). The remaining 13 patients with 26 wounds had WF, requiring additional debridements after wound closure was attempted (figure IB). Patient serum and wound effluent level of cytokines varied significantly from the day of admission to the day of DWC. These significant variations were associated with wound failure (WF, dehiscing) or successful healing (H) and detailed in Figures 1C, ID, and 2A-D and in Table 1.

Table 1. Univariate p-values.

Serum Effluent

All Time Points Final Time Point All Time Points Final Time Point

All Healed Dehisced All Healed Dehisced All Healed Dehisced All Healed Dehisced

EGF 0.470 0.765 0.235 0.079 0.961 0.039 0.115 0.943 0.057 0.707 0.650 0.353

Eotaxin 0.527 0.737 0.264 0.568 0.284 0.718 0.695 0.347 0.653 0.585 0.292 0.710

FGFBasic 0.006 0.003 0.997 0.058 0.029 0.972 0.012 0.994 0.006 0.102 0.950 0.051

GCSF 0.030 0.015 0.985 0.507 0.254 0.749 0.014 0.993 0.007 0.054 0.974 0.027

GMCSF 0.159 0.080 0.921 0.316 0.158 0.844 0.003 0.001 0.999 0.038 0.019 0.981

HGF 0.004 0.998 0.002 0.003 0.999 0.001 0.794 0.604 0.397 0.905 0.453 0.550

IFNa 0.084 0.958 0.042 0.630 0.688 0.315 0.008 0.996 0.004 0.454 0.775 0.227

IFNy 0.012 0.994 0.006 0.217 0.893 0.109 0.003 0.999 0.001 0.890 0.558 0.445

IL10 0.001 1.000 0.000 0.139 0.931 0.070 0.011 0.994 0.006 0.344 0.830 0.172

IL12 0.000 0.000 1.000 0.026 0.013 0.987 0.077 0.038 0.962 0.618 0.309 0.694

IL13 0.236 0.118 0.882 0.728 0.639 0.364 0.069 0.034 0.966 0.030 0.015 0.985

IL15 0.003 0.999 0.001 0.111 0.946 0.055 0.078 0.961 0.039 0.604 0.302 0.701

IL17 0.000 0.000 1.000 0.089 0.045 0.956 0.013 0.007 0.993 0.104 0.052 0.949

ILla 0.477 0.762 0.239 0.704 0.651 0.352 0.281 0.860 0.141 0.440 0.782 0.220

IL1B 0.291 0.146 0.855 0.673 0.336 0.666 0.076 0.962 0.038 0.324 0.840 0.162

IL2 0.006 0.003 0.997 0.492 0.246 0.756 0.000 0.000 1.000 0.026 0.013 0.987

IL2R 0.002 0.999 0.001 0.296 0.854 0.148 0.001 1.000 0.000 0.289 0.857 0.144

IL3 0.952 0.476 0.524 0.845 0.422 0.580 0.654 0.327 0.674 0.802 0.602 0.401

IL4 0.051 0.026 0.974 0.785 0.392 0.610 0.000 0.000 1.000 0.006 0.003 0.997

IL5 0.462 0.769 0.231 0.637 0.684 0.319 0.728 0.364 0.637 0.001 0.000 1.000

IL6 0.000 1.000 0.000 0.014 0.993 0.007 0.000 1.000 0.000 0.106 0.948 0.053

IL7 0.000 1.000 0.000 0.010 0.995 0.005 0.000 1.000 0.000 0.390 0.807 0.195

IL8 0.291 0.145 0.855 0.444 0.780 0.222 0.061 0.030 0.970 0.271 0.135 0.866

IL1RA 0.069 0.034 0.966 0.152 0.076 0.925 0.985 0.508 0.493 0.379 0.190 0.812

IP10 0.000 0.000 1.000 0.003 0.002 0.998 0.005 0.002 0.998 0.015 0.007 0.993

MCP1 0.003 0.999 0.001 0.851 0.577 0.425 0.026 0.987 0.013 0.179 0.912 0.090

MIG 0.440 0.220 0.780 0.076 0.038 0.963 0.841 0.580 0.420 0.006 0.003 0.997

MlPla 0.037 0.018 0.982 0.360 0.180 0.822 0.958 0.479 0.522 0.384 0.810 0.192

MIP1B 0.223 0.112 0.889 0.444 0.222 0.780 0.965 0.518 0.483 0.742 0.632 0.371

RANTES 0.000 1.000 0.000 0.005 0.998 0.003 0.465 0.768 0.232 0.537 0.734 0.268

TNFa 0.002 0.001 0.999 0.181 0.091 0.911 0.058 0.029 0.971 0.728 0.364 0.639

VEGF 0.010 0.995 0.005 0.042 0.979 0.021 0.064 0.968 0.032 0.384 0.810 0.192

[0060] Planned timing for DWC varied according to wound type with amputation wounds being closed the latest after injury (Figure 3A). Wound types were further evaluated to determine if any category, STI, OF or A, occurred as a single wound or if it occurred as one of multiple wounds by patient (p<0.01) (Figure 3B). When comparing each isolated wound type, amputations were more likely to occur in patients with multiple extremity wounds (OR= 5.6, 95% CI 2.3 to 13.5) and open fractures were more likely to be single wounds (OR= 4.4, 95% CI 1.7 to 11). In patients with multiple wounds, DWC occurred on later dates after the day of injury (Figure 3C), although surgeons were less likely to accomplish successful closure in patients with multiple wounds (Figure 3D). These wounds had increased odds of failing to heal properly and require further surgical debridements after DWC (OR= 17.4, 95% CI 2.8 to 184.7).

[0061] After model development using Bayes classifiers to identify cytokine association with the time of DWC, a BBN model was trained based on the cytokines identified and clinical variables previously associated with the time of DWC. These cytokines and clinical variables were: serum levels of I L-6, I L-7, I L-8, I L-10, and VEGF, effluent levels of I L-lra, I L-5, I L-7, I FN-y, FGF-basic, MIG, and ΜΙ Ρ-Ιβ, injury type, occurrence of multiple wounds, number of days from injury and previous surgical debridements after admission. After five-fold cross-validation, this BBN model (Figure 4) successfully estimated the time for immediate DWC and the need for additional surgical debridements. The time of DWC was estimated as the need for immediate wound closure

(AUC=0.82; 95% CI, 0.79-0.86) and the requirement of additional debridements before successful wound closure as: one (AUC=0.61; 95% CI, 0.507-0.72), two (AUC=0.66; 95% CI, 0.52-0.79), and three or more (AUC=0.7; 95% CI, 0.67-0.73) surgical debridements at WRN M MC (Figure 5). The inclusion of additional variables such as ISS, APACH E I I, and wound CC status did not improve the performance of this DWC model. The level of all cytokines present in the current model was associated with the number of surgical debridements needed for successful DWC as shown by each correlation coefficient (Table 2). Of those, serum cytokines had positive correlation and effluent cytokines, mostly negative correlation, with exception of bFGF. As a first degree associate, the level of I L-5 also had the strongest relationship and coefficient of -0.54 (95% CI, -0.66 to -0.44).

Table 2: Spearman's rank correlation of BBN model cytokines with the number of necessary surgical debridements for successful wound closure. Statistical significant difference (p<0.01), number of serum and effluent samples (N), correlation coefficient (rs) and 95% confidence interval (CI) are shown as indicated.

Cytokines (all, p<0.01) r_s (-1 to 1) 95%CI

Serum:

I L-6 0.38 0.28 to 0.48

I L-7 0.20 0.08 to 0.30

I L-8 0.32 0.22 to 0.42

IL-10 0.32 0.21 to 0.42

VEG F 0.30 0.19 to 0.40

Effluent:

IL-1RA -0.46 -0.55 to -0.35

I L-5 -0.54 -0.62 to -0.44

I L-7 -0.21 -0.33 to -0.09

I FN-γ -0.39 -0.49 to -0.28

bFGF 0.24 0.12 to 0.36

M IG -0.27 -0.39 to -0.15

ΜΙ Ρ-1β -0.31 -0.42 to -0.19 [0062] Considering DWC as the event of interest, Kaplan-Meier curve analysis revealed successful healing (p=0.02), injury type (p<0.01), development of CC (p<0.01), and the occurrence of multiple extremity wounds (p<0.01) were all variables associated with the time of DWC. Surgeons were more likely to incorrectly plan successful DWC in patients with multiple extremity wounds, as these wounds had increased odds of failed closure, OR=17.45 (2.8 to 184).

[0063] The timing of delayed wound closure was investigated to determine if it could be estimated by clinical variables and the local and systemic cytokine response of combat-related extremity wounds. This estimation considered each wound independently, which could be one, two or, maximally, three per patient. Only wounds that closed successfully after a series of subsequent surgical debridements were evaluated to estimate the timing of closure. By considering only these wounds, possible variations in the level of inflammatory cytokines from the day of admission to the day of wound closure were used as an objective measurement of wound healing. Following standard treatment guidelines, surgeons successfully achieved wound closure at the initial attempt in approximately 77% of the 116 extremity wounds evaluated in the present study. Wounds with successful healing were, in general, closed at later dates.

[0064] Local and systemic cytokines levels were initially evaluated by univariate analysis, comparing the day of admission, as baseline, with the day of wound closure. This analysis showed that patients with wound failure and successful closure had significant variations in the level of cytokines during treatment. Patients with WF demonstrated, in general, cytokine levels more similar to baseline values on admission to the hospital. All groups of cytokines, interleukins, growth factors and chemokines have demonstrated variations in the local and systemic level during treatment from admission to DWC. In serum samples, the only cytokine with lower concentrations in both successfully healed and failed wounds at DWC was IP-10. In addition to attracting inflammatory cells, this chemokine induced by IFN-y is thought to also inhibit angiogenesis in vivo (Angiolillo, Human interferon-inducible protein 10 is a potent inhibitor of angiogenesis in vivo. J Exp Med. 1995; 182, 155-62). Successful wound closure occurred in wounds with an isolated increase in local effluent levels of pro-inflammatory cytokines such as IL-8 and IL-17 and was associated with decreased level of anti-inflammatory IL-10. This relationship between IP-10, VEGF, IL-8, IL-17 and IL-10 exemplifies some of the functional variations in the level of isolated cytokines.

[0065] A combined multivariate analysis approach using machine learning methods was performed to estimate the occurrence of a successful outcome and the need for further surgical treatment. First, a Naive Bayes classifier was used to identify cytokines associated with the timing of DWC and, later, train a Bayesian belief network. To identify additional categorical variables associated with the timing of DWC for the BBN model, survival curve analysis demonstrated injury type, the occurrence of multiple or single wounds per patient and the development of CC as possible candidate variables aiming at improving the its performance. However, the inclusion of data regarding the occurrence of wound CC did not improve the performance of the BBN model.

[0066] As a first-degree associate, the local effluent level of IL-5 demonstrated the strongest association with the estimated necessary number of debridements until successful DWC. This cytokine is thought to play an important role in eosinophil recruitment during the inflammatory and granulation phases of wound repair. As demonstrated in the BBN model, variations in the local concentration of IL-5 in association with other cytokines such IL-8, IL-6 and IFN-y promote healing and contribute to estimate the timing of successful closure. Both the initial univariate analysis comparing the level of cytokines on admission and at wound closure, in addition to the correlation analysis of cytokines involved in the BBN model showed, in general, systemic level of cytokines decrease and local level of cytokines increase approaching the timing of successful wound closure. This is associated with more local production of cytokines from cells already outside the blood stream. In cross validation, the BBN model demonstrated an improved performance estimating the need for immediate closure and the need of three or more surgical debridements before successful wound closure.

[0067] The optimal timing for DWC and the need of further surgical debridements in the treatment of combat-related extremity wounds can be determined by cytokine response and clinical variables. These results form the basis of a novel clinical decision support tool and assist surgeons in determining the optimal time to close wounds while also aiding in estimating the necessary number of future surgeries. This allows the surgeon to plan the DWC as early as possible and to estimate the duration of the surgical treatment necessary to successfully close each extremity wound individually.

Claims

Claims

1. A method of treating an open wound in a human subject comprising:

(a) obtaining a first blood, serum or plasma sample from the human subject,

(b) determining levels of one or more cytokines selected from interleukin-6 (I L-6), interleukin-7 (I L-7), interleukin-8 (I L-8), interleukin-10 (I L-10) and vascular endothelial growth factor (VEGF) in the first blood, serum or plasma sample, and

(c) surgically closing the wound within 24 hours in human subjects with cytokine levels about the same as human subjects with successful wound closure.

2. The method of claim 1 further comprising:

(d) obtaining a second blood, serum or plasma sample from the human subject at least 24 hours after obtaining the first blood, serum or plasma sample, and

(e) determining levels of one or more cytokines selected from I L-6, I L-7, I L-8, IL-10 and VEGF in the second blood, serum or plasma sample.

3. A method of treating an open wound in a human subject comprising:

(a) obtaining a first wound effluent sample from the human subject,

(b) determining levels of one or more cytokines selected from interleukin-1 receptor agonist (I L-lra), interleukin-5 (I L-5), interleukin-7 (I L-7), basic fibroblast growth factor (FGF-basic), interferon gamma (I FG), monokine-induced gamma interferon (MIG), and macrophage inflammatory protein-1 (M I P-1) in the first wound effluent sample, and

(c) surgically closing the wound within 24 hours in human subjects with cytokine levels about the same as human subjects with successful wound closure.

4. The method of claim 3 further comprising:

(d) obtaining a second wound effluent sample from the human su bject at least 24 hours after obtaining the first blood, serum or plasma sample, and

(e) determining levels of one or more cytokines selected from I L-lra, I L-5, IL-7, FGF-basic, I FG, M IG, and MI P-1 in the second wound effluent sample.

5. The method of claim 1 further comprising obtaining a first wound effluent sample from the human subject, and determining levels of one or more cytokines selected from I L-lra, I L-5, I L-7, FGF-basic, IFG, MIG, and M IP-1 in the first wound effluent sample.

6. The method of claim 9 further comprising obtaining a second blood, serum or plasma sample from the human subject at least 24 hours after obtaining the first blood, serum or plasma sample, obtaining a second wound effluent sample from the human subject, determining levels of one or more cytokines selected from IL-6, IL-7, IL-8, IL-10 and VEGF in the second blood, serum or plasma sample, and determining levels of one or more cytokines selected from IL-lra, IL-5, IL-7, FGF-basic, IFG, MIG, and MIP-1 in the second wound effluent sample.

7. The method of any previous claim further comprising debriding the wound.

8. The method of any previous claim further comprising waiting at least two days before

surgically closing the wound.

9. The method of any previous claim further comprising determining the number of open wounds in the human subject.

10. The method of any previous claim further comprising determining the number of days after the human subject was wounded.

11. The method of any previous claim further comprising determining the number of times the wound has been debrided.

12. The method of claims 2, 4, or 6 further comprising waiting 48 to 72 hours after obtaining the first sample.

13. The method of claims 2, 4, 6 or 12 further comprising debriding the wound.

14. The method of claims 2, 4, 6 or 12-13 wherein the wound is surgically closed after

determining cytokine levels in the second sample.

15. The method of any previous claim wherein the human subject has a trauma wound.

16. The method of any previous claim wherein the open wound is an extremity wound.

17. A method of determining whether to surgically close an open wound comprising:

(a) obtaining a blood, serum or plasma sample from the human subject; and

(b) measuring levels of one or more cytokines selected from IL-6, IL-7, IL-8, IL-10 and VEGF in the blood, serum or plasma sample;

wherein cytokine levels about the same as human subjects with successful wound closure indicate that the wound should be closed within 24 hours.

18. A method of determining whether to surgically close an open wound comprising:

(a) obtaining a wound sample from the human subject; and

(b) measuring the levels of IL-lra, IL-5, IL-7, FGF-basic, IFG, MIG, and MIP-1 in the wound effluent sample;

wherein cytokine levels about the same as human subjects with successful wound closure indicate that the wound should be closed within 24 hours.

19. The method of claim 17 further comprising:

(d) obtaining a wound sample from the human subject; and

(e) measuring the levels of IL-lra, IL-5, IL-7, FGF-basic, IFG, MIG, and MIP-1 in the wound effluent sample.

20. The method of any of claims 17-19 wherein the wound is an extremity wound.

21. The method of any of claims 17-20 wherein the open wound is a traumatic wound.

22. The method of any of claims 17-21 further comprising determining the number of open wounds in the human subject.

23. The method of any of claims 17-22 further comprising determining the number of days after the human subject was injured.

24. The method of any of claims 17-23 further comprising determining the number of times the wound has been debrided.