Precision medicine has not been widely adopted in surgical and trauma care. Trauma care offers a unique opportunity for precision approaches. Researchers have shown that individual differences in the response to injury, most notably in the immune response, have significant effects on outcomes.1–3 In addition, chronic changes in individual immune and metabolic adaptations to injury are clearly associated with poor long-term function and impaired healing.4,5 Orthopaedic trauma patients are at particular risk of complications. Patients sustaining severe injury have frequent multisystem issues with extensive skin, muscle, and bone wounds all at risk of complicated healing. Surgical intervention is the foundation of treatment of these composite wounds; however, choices and timing of surgical interventions often present challenging decisions to the treating surgeon. These decisions are often guided by collective experience of the surgical team, and the interventions are chosen anecdotally. The consequences can be devastating with nonunions, bone and soft-tissue infections, and markedly weak and dysfunctional limbs. Advances in computational data modeling offer novel opportunities to individualize treatment decisions. In this symposium, we discuss 3 topics that are core to the mission of the orthopaedic trauma surgeon including managing severe open musculoskeletal wounds, staging interventions in multiply injured patients with major skeletal trauma, and nonunions. Advances in precision and computational approaches to improve management of these challenging conditions are emphasized.
PRECISION METHODS FOR WOUND MANAGEMENT
The Uniformed Services University Surgical Critical Care Initiative (SC2i) was established in 2013 to develop Clinical Decision Support Tools (CDSTs) for acute trauma care. A consortium of 7 federal and nonfederal entities (Uniformed Services University, Walter Reed National Military Medical Center, Naval Medical Research Center, Henry M. Jackson Foundation, Duke University, Emory University, and DecisionQ Corporation) form the SC2i which is the Department of Defense's premier precision medicine center for acute and trauma care.
In its mission to bring highly personalized diagnoses and therapeutic interventions to its wounded service members, the SC2i incorporates best practices in data science to ensure high-quality data are available for all of its translational and clinical research. Thus far, the SC2i has enrolled greater than 1600 patients, representing 57,800 biobanked specimens and 20+ million data elements. Clinical and bioassay instrumentation data are aggregated across all consortium partners in a Central Data Repository on Amazon Web Services (GovCloud). The data aggregation process involves site and consortium data managers, all focused on data standardization and quality.
SC2i leverage patients' transcriptomic, proteomic, bacteriological, and clinical data, as well as advanced machine learning techniques, to develop CDSTs for conditions associated with a high risk of morbidity or mortality. The SC2i has released 3 CDSTs to date for the activation of a massive transfusion protocol, the prediction of the onset of sepsis, and the prediction of the development of invasive fungal infections. There are CDSTs in development thus far by SC2i to predict the development of pneumonia, bacteremia, venous thromboembolisms, acute kidney injury, and delayed closure of extremity wounds. Currently, the time of closure of combat-related extremity wounds is planned based on the extent of injury, signs of necrosis, and resolution of possible infections but is often anecdotally decided by the surgeon. Approximately 80% of these wounds are closed successfully and heal without subsequent complications. Healing is dependent on a suitable cytokine response. The association of inflammatory cytokines, chemokines, and growth factors with successful healing and proper timing of surgical definitive closure is important in model training to develop CDSTs to aid on the treatment planning of these wounds. The wound closure CDST model is informed by serum-based and wound exudate-based cytokine concentrations in conjunction with demographic data. Based on these findings, SC2i investigators have trained models to predict successful definitive closure, the necessary number of debridement surgeries to treat extremity wounds, wound failure after development of critical colonization, and the development of heterotopic ossification, all with mean area under the curve in receiver operating curve curves equal or higher than 80% (Table 1).
While the main goal is to produce CDSTs, this research has elucidated potential avenues for research into the functional mechanisms underlying various complications seen in the surgical critical care setting. A key player in precision medicine for surgical critical care, SC2i uses high-quality data, standardized across multiple institutions, to develop clinically relevant CDSTs expected to identify favorable interventions, improve outcomes and resource utilization in both civilian and military health systems.
PRECISION METHODS TO STAGE ORTHOPAEDIC INTERVENTIONS
“Precision medicine” is defined by the National Institute of Health as an emerging approach for disease prevention and treatment that takes into account an individual's genetic variation, environment, and lifestyle. Technologic advances that facilitate inexpensive and expeditious genome sequencing have led to the development of precision treatment approaches in the field of oncology and revolutionized cancer treatment. There is an increased emphasis placed on precision approaches in many disciplines including cardiology, endocrinology, and rheumatology. A personalized, or precision, approach to trauma has lagged behind other specialties.6 Orthopaedic trauma surgery, in particular, has focused research and treatment toward population-based approaches.
Optimal timing of definitive fixation for pelvis and long bone fractures remains controversial. In the 1980s and early 1990s, Pape et al7 demonstrated the beneficial systemic effects of temporary fixation of long bone fractures [“damage control orthopaedics” (DCO)] before definitive fixation in vulnerable patient populations. Patients who were hemodynamically unstable, acidotic, coagulopathic, or sustained a severe head or chest injury were less likely to succumb to subsequent lung injury and systemic complications when initially treated through temporizing measures for long bone fractures.7,8 Clinicians have traditionally advocated avoiding definitive fixation of long bone fractures on days 2–5 in a potential “second-hit” window that may trigger an already primed immunologic system leading to a deleterious systemic inflammatory response. Researchers have recently called into question the “second-hit” model of inflammation and have proposed an alternative model of postinjury inflammation characterized by an upregulation of the innate immune response and a simultaneous suppression of the adaptive immune response that can persist and lead to long-term immune dysregulation.4,5,9 Patient-specific differences in proinflammatory and anti-inflammatory responses after injury likely explain discrepancies in outcomes to interventions in multiply injured patients.
Recent evidence also suggests that most appropriately resuscitated trauma patients without severe head or chest injury benefit from early fixation of axially unstable fractures.10 This “early appropriate care” (EAC) model has demonstrated that physiologically stable trauma patients undergoing early definitive fixation have fewer systemic complications and a shorter hospital stay.11 Current evidence to guide initial and staged orthopaedic interventions relies on anatomic injury description, resuscitation markers, and hemodynamic status. Measures of acidosis (pH, base deficit, and lactate), vital sign parameters, and injury patterns (injury severity score and anatomical description) used to direct interventions are rooted in investigations of large trauma populations. Although population-based algorithms derived to guide fracture interventions have proven useful and have greatly improved care of multiply injured patients with orthopaedic injuries, there has been recent attention in precision approaches to this unique population. Specifically, there is growing interest in the field of trauma immunology and utilization of biomarkers to identify patients at risk of adverse outcomes very early after the onset of disease. Trauma is one of very few conditions in which the exact time of disease onset can be determined. The immune-mediated response to injury, measured by immunologic biomarkers, has been shown to have a significant impact on outcome in trauma patients.3–5,12 The authors have recently demonstrated associations between elevated immunologic mediators within 24 hours of injury and higher levels of organ dysfunction in orthopaedic trauma patients.13 Subsequent work has focused on evaluation of the varied immunologic response among patients treated with early (≤48 hours) versus late (>48 hours) definitive fixation of femur, pelvis, and acetabular fractures.
We enrolled 61 polytraumatized patients with orthopaedic injuries into a prospective, observational study that examined the association between immunologic biomarkers and short-term clinical outcome. A subgroup analysis of 19 patients with operative femur fractures was subsequently conducted. Ten patients underwent early (≤48 hours) definitive fixation of a femur fracture (EAC), and 9 patients underwent late (>48 hours) definitive fixation after temporizing external fixation or skeletal traction. Blood was drawn on all patients 3 times within the first 24 hours of injury and daily while in the intensive care unit. Plasma analyses were performed using multiplex ELISA methods to quantify 20 immunologic mediators. The following inflammatory mediators were measured: interleukin (IL)-1β, IL-1 receptor antagonist (IL-1RA), soluble IL-2 receptor-α (sIL-2Rα), IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-17A, IL-17E/IL-25, IL-21, IL-22, IL-23, IL-33, interferon gamma-induced protein 10 (IP-10), monokine induced by interferon gamma, monocyte chemoattractant protein-1, and high mobility group box 1. Clinical data were gathered in prospective fashion including demographics, injury characteristics, resource utilization, and short-term outcome data (infection and organ dysfunction).
There was no difference between the groups with respect to age, sex, or body mass index. As expected, the DCO group had a longer hospital length of stay (mean 27 days vs. 17, P = 0.03) and greater transfusion requirements (mean 20U vs. 4U of packed red blood cells, P = 0.03), In addition, DCO patients trended toward a higher injury severity (injury severity score of 40 vs. 30, P = 0.11), longer intensive care unit length of stay (mean 15 days vs. 9, P = 0.15), and longer time on mechanical ventilation (mean 11 days vs. 6, P = 0.15) compared with the EAC group. Interestingly, there were no differences observed in biomarker concentrations between the DCO and EAC groups at time point 0, 24 hours, and 48 hours after injury. Several scattered differences were observed between the groups on day 3 (mean IL-22 DCO 360.6 vs. EAC 827.2, P < 0.05; mean sIL-2RA DCO 1669.6 vs. EAC 1087.5, P < 0.05), day 4 (mean IL-10 DCO 41.5 vs. EAC 13.0, P < 0.05), and day 5 (mean IL-8 DCO 54.1 vs. EAC 24.0, P < 0.05), but no wide scale differences occurred at any time points. In this small sample, we observed no proinflammatory spike seen at the time of a second “definitive fixation” surgery in the DCO group between days 2 and 5, or any discernible perturbation in cytokines after femoral fixation in the EAC group.
Analysis of another subgroup of 41 patients with pelvic and acetabular fractures from the same prospective cohort described above was performed to further explore the association between surgical timing and immunologic activity. We hypothesized that patients undergoing pelvic and acetabular surgery after 48 hours (LATE, n = 9) within the second-hit window after injury would have significant elevations in biomarkers perioperatively compared wih patients who had surgery before 48 hours (EARLY, n = 8) after injury. A third nonoperatively treated group served as a control (n = 24). Similar to the femoral fixation group, there were no differences in inflammatory mediator concentrations after definitive pelvis and acetabular surgery between the EARLY and LATE groups. There was no surge in any proinflammatory mediator in patients having surgery in the day 2–5 window. Although no conclusions can be drawn from this investigation given the small sample size and limited number of mediators studied, no association was observed with respect to the timing of pelvic and acetabular surgery in polytraumatized patients and the subsequent immunologic response.
In both subcohort analyses, a surgically induced “second-hit” inflammatory response was not observed. Larger, prospective studies are necessary to quantify the impact of staged fracture interventions on the systemic inflammatory response and relevant clinical outcomes. In addition, these limited data highlight shortcomings in reductionist approaches to understanding immunologic perturbations resulting from interventions in trauma patients. Advances in computational methods to explore biomarker orchestration in trauma patients are likely necessary to understand how interventions affect individualized immunologic response.12
PRECISION METHODS FOR NONUNIONS: A BLOOD-BASED EARLY DIAGNOSTIC TEST
Three biologic outcomes occur after fracture—normal healing (union), slow healing (delayed union), or failed healing (nonunion).14 Abnormal fracture healing (delayed union and nonunion) can result in significant deficits in function and have a dramatic effect on clinical and patient-reported outcomes. Several publications have described or pursued the need for a definitive prognostic tool and effective treatments for delayed union and nonunion.14–18 There is evidence that most, if not all, risk factors associated with delayed union and nonunion are present at the time of injury and/or index presentation.16,18,19 The relative importance of each global risk factor, however, remains unknown with respect to guiding specific clinical decisions from one patient to the next.15,18,20 The need for confident prognosis or earlier diagnosis of delayed union or nonunion continues to be at the forefront of modern fracture care.15,17 Similar uncertainty in cancer or cardiac disease would not be tolerated.
Currently, standard of care results in definitive diagnosis of fracture nonunion being delayed several months or waiting until catastrophic failure. From the perspective of enrolling a potential nonunion patient in a clinical trial, the timeline to diagnosis is more firmly codified to 6–9 months after index treatment.21 The ability to confirm early prognosis of nonunion would (1) liberate the surgeon to engage in more aggressive initial treatments, (2) permit earlier revision surgeries, (3) provide patients with necessary information to adjust their care plans and their expectations, and (4) expand opportunities to trial interventions to the appropriate target population. To date, attempts have been used to leverage tissue samples and clinical, radiographic, and demographic information to stratify nonunion risk. These approaches include analysis of epidemiologic, clinical, biochemical, bone turnover, genetic, radiologic, and computational data.16,19,22–24 Currently, no diagnostic test exists to definitively discern whether a patient will heal normally, experience delayed union, or nonunion.
Our work has focused diagnosing nonunion in the acute postinjury phase (<7 days after injury). Peripheral blood was used as the surrogate tissue and both miRNA and mRNA as the target analytes. After receiving approval from the University of Pennsylvania Institutional Review Board, adult subjects (>18 years old) were enrolled into 3 cohorts including healthy volunteers (n = 35), acutely injured subjects (AIS) with new fractures (n = 129), and subjects with chronic nonunion (NU) (n = 12). Blood was collected on a cohort-specific schedule. NU provided specimens before and after revision surgeries, and AIS provided specimens throughout their postinjury care that were then grouped according to 4 time windows. Initially, microarrays were performed for high-level screening across the 3 cohorts to identify RNA expression changes unique to AIS.25 real-time quantitative polymerase chain reaction was then performed on selected RNAs (204 mRNA and 25 miRNA) to confirm the direction and magnitude of individual RNA performance.26 Regression analysis was used to model the within AIS cohort outcome correlations to single and multiple biomarker expression. Receiver operator curves were generated using 85% of the data and verified using the remaining 15%.
This single-center pilot study revealed a short list of 45 mRNA/miRNA that either alone or in pairs resulted in at least a 75% positive predictive value (PPV), P < 0.05, for identifying eventual nonunion or delayed union. Specific combinations of 2 or 3 biomarkers resulted in >90% PPV, P < 0.05. The addition of a fourth biomarker to the models did not improve PPV. Blood RNA profiles also permitted the distinction across healing outcomes within the AIS cohort at early time points after injury and with desirably high confidence (75%–90% PPV). The commitment to the healing process occurs within the earliest sampling window (≤7 days) and maintains through the period of observation in this study population, but the specific RNAs that signal the nonunion outcome are not identical from one time window to the next, which is consistent with the progressive biological processes involved in fracture healing. Enrollment, sample collection, analysis, additional model refinement, and confirmatory studies are ongoing to finalize a diagnostic test for future clinical trial validation.27 However, these pilot data do reveal the opportunity of using peripheral blood after orthopaedic trauma in predicting fracture healing in a personalized, patient-specific manner.
Precision methods for diagnosis and intervention will continue to progress in all fields of medicine and surgery including orthopaedic surgery. Trauma-relevant phenotypes including traumatic extremity wounds, nonunions, and multiply injured patients with operative skeletal injuries are presented as examples in which precision methods have been initially interrogated. These approaches offer great potential to improve outcomes in orthopaedic trauma patients with severe extremity injuries.
1. Lord JM, Midwinter MJ, Chen YF, et al. The systemic immune response to trauma: an overview of pathophysiology and treatment. Lancet. 2014;384:1455–1465.
2. Mi Q, Constantine G, Ziraldo C, et al. A dynamic view of trauma/hemorrhage-induced inflammation in mice: principal drivers and networks. PLoS One. 2011;6:e19424.
3. Namas RA, Vodovotz Y, Almahmoud K, et al. Temporal patterns of circulating inflammation biomarker networks differentiate susceptibility to nosocomial infection following blunt trauma in humans. Ann Surg. 2016;263:191–198.
4. Gentile LF, Cuenca AG, Efron PA, et al. Persistent inflammation and immunosuppression: a common syndrome and new horizon for surgical intensive care. J Trauma Acute Care Surg. 2012;72:1491–1501.
5. Vanzant EL, Lopez CM, Ozrazgat-Balanti T, et al. Persistent inflammation, immunosuppression, and catabolism syndrome after severe blunt trauma. J Trauma Acute Care Surg. 2014;76:21–29; discussion 29–30.
6. Buchman TG, Billiar TR, Elster E, et al. Precision medicine
for critical illness and injury. Crit Care Med. 2016;44:1635–1638.
7. Pape HC, Hildebrand F, Pertschy S, et al. Changes in the management of femoral shaft fractures in polytrauma patients: from early total care to damage control orthopedic surgery. J Trauma. 2002;53:452–461; discussion 461–462.
8. Pape HC, Giannoudis PV, Krettek C, et al. Timing of fixation of major fractures in blunt polytrauma: role of conventional indicators in clinical decision making. J Orthop Trauma. 2005;19:551–562.
9. Xiao W, Mindrinos MN, Seok J, et al. A genomic storm in critically injured humans. J Exp Med. 2011;208:2581–2590.
10. Vallier HA, Super DM, Moore TA, et al. Do patients with multiple system injury benefit from early fixation of unstable axial fractures? The effects of timing of surgery on initial hospital course. J Orthop Trauma. 2013;27:405–412.
11. Nahm NJ, Como JJ, Wilber JH, et al. Early appropriate care: definitive stabilization of femoral fractures within 24 hours of injury is safe in most patients with multiple injuries. J Trauma. 2011;71:175–185.
12. Namas RA, Almahmoud K, Mi Q, et al. Individual-specific principal component analysis of circulating inflammatory mediators predicts early organ dysfunction in trauma patients. J Crit Care. 2016;36:146–153.
13. Gaski GE, Metzger C, Wessel R, et al. The early immunologic response in multiply injured patients with orthopaedic injuries is associated with organ dysfunction. J Orthop Trauma. 2019.
14. Estes WL. A study of the cause of delayed union and non-union in fractures of the long bones. Ann Surg. 1920;71:40–46.
15. O'Halloran K, Coale M, Costales T, et al. Will my tibial fracture heal? Predicting nonunion
at the time of definitive fixation based on commonly available variables. Clin Orthop Relat Res. 2016;474:385–395.
16. Zura R, Xiong Z, Einhorn T, et al. Epidemiology of fracture nonunion
in 18 human bones. JAMA Surg. 2016;151:e162775.
17. Simpson A. The forgotten phase of fracture healing: the need to predict nonunion
. Bone Joint Res. 2017;6:610–611.
18. Perumal V, Roberts CS. (ii) Factors contributing to non-union of fractures. Curr Orthop. 2007;21:258–261.
19. Axelrad TW, Einhorn TA. Use of clinical assessment tools in the evaluation of fracture healing. Injury. 2011;42:301–305.
20. Cunningham BP, Brazina S, Morshed S, et al. Fracture healing: a review of clinical, imaging and laboratory diagnostic options. Injury. 2017;48(suppl 1):S69–S75.
21. Guidance Document for Industry and CDRH Staff for the Preparation of Investigational Device Exemptions and Premarket Approval Applications for Bone Growth Stimulator Devices (Docket No. 98D-0238). Rockville, MA: U.S. Department of Health and Human Services, Food and Drug Administration, Center for Devices and Radiological Health, Division of General and Restorative Devices, Office of Device Evaluation; 1998.
22. Dimitriou R, Carr IM, West RM, et al. Genetic predisposition to fracture non-union: a case control study of a preliminary single nucleotide polymorphisms analysis of the BMP pathway. BMC Musculoskelet Disord. 2011;12:44.
23. Ivaska KK, Gerdhem P, Akesson K, et al. Effect of fracture on bone turnover markers: a longitudinal study comparing marker levels before and after injury in 113 elderly women. J Bone Miner Res. 2007;22:1155–1164.
24. Anderson DD, Thomas TP, Campos Marin A, et al. Computational techniques for the assessment of fracture repair. Injury. 2014;45(suppl 2):S23–S31.
25. Mehta S, Horan AD. 809094 biomarker identification in fracture healing (experiment #5201-5202 jun 2015) deposited in NCBI's gene expression omnibus and accessible through GEO series accession number GSE93387. Available at: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE93387
. Accessed January 11, 2017.
26. Baldwin DA, Horan AD, Hesketh PJ, et al. Combined RT-qPCR of mRNA and microRNA targets within one fluidigm integrated fluidic circuit. J Biomol Tech. 2016;27:75–83.
27. Horan AD, Mehta S, Baldwin DA. (WO2016126844) Novel Methods for Early Identification of Bone Healing Ability in Injured patients, International Application No.:PCT/US2016/016404, International Filing Date: 03Feb2016 Patent Pending. (Priority Data: 62/231,935 03Feb2015 US and 62/283,443 01Sep2015 US). Philadelphia, PA: The Trustees of the University of Pennsylvania [US/US]; 2017.