BACKGROUND AND RATIONALE
Acute compartment syndrome (ACS) occurs in as many as 11%–18% of high-energy tibia fractures1,2 and remains a diagnostic challenge, a source of morbidity for trauma patients, and adds significant costs to caring for the injured patient.3–6 There are currently no validated clinical criteria that reliably identify when ACS is present. Instead, clinicians rule out ACS based on the absence of concerning clinical findings and, in some cases, by demonstrating that tissue perfusion pressure (PP) is “safe” (PP ≥30 mm Hg).1 However, for both clinical findings and isolated measurements of intramuscular pressure (IMP) or PP, there are few prospective clinical studies assessing their sensitivity or positive predictive value, limiting the application of their use to reliably diagnose ACS.7,8 What little data exist raises concerns for the efficacy of these tests, both in the reliability of the test itself,9 and the relationship between the results of the test and the diagnosis of ACS.7,8,10
ACS remains a challenge to diagnose in patients with extremity trauma, and because of the devastating nature of the consequences of missed compartment syndrome, it is a common source of litigation in civilian practice.11 Patients with concerning clinical findings and/or pressure measurements suspicious for ACS are treated with urgent surgical fasciotomy12 which immediately reduces IMP and restores myoneural perfusion.7 It is likely, however, that some patients undergoing fasciotomy do not have compartment syndrome and are receiving unnecessary and markedly morbid surgery. Although fasciotomies introduce complications and morbidity, the practice is accepted because the morbidity of delayed fasciotomy or missed diagnosis of ACS can be greater. Cases of ACS that are not treated with timely fasciotomy are associated with muscle necrosis, which in the short term may cause rhabdomyolysis and renal failure, and in the long term may result in ischemic contracture of the involved compartment and permanent functional deficit.13 In 2 series of patients, delayed and incomplete fasciotomy were associated with increased morbidity and mortality, and a high rate of limb amputation.14,15
Because the current management of ACS is based on subjective clinical assessments, it is not surprising that the rate of fasciotomy varies significantly among clinicians.10 Less often discussed is the fact that all research published to date on ACS is affected by the lack of a strict gold standard for diagnosis. Lacking such a standard, the performance of fasciotomy has been used in past studies as a proxy for the presence of ACS, where patients who undergo fasciotomy are considered to have had ACS, despite the fact that an unknown proportion of them did not actually have the condition. This lack of a strict definition of ACS compromises understanding of the condition, including its diagnosis and treatment. The development of a decision rule for ACS could by itself reduce the incidence of fasciotomy simply by lessening the number of false-positive diagnoses. It would also guard against the potential for a missed (or delayed) diagnosis of ACS. Such a decision rule would also facilitate further research on potential therapies for ACS by providing a set of standardized criteria for the diagnosis of ACS.
No one has carefully mapped the relationship between clinical signs, objective, continuous measurements of leg perfusion or IMP and a diagnosis of ACS. Continuous pressure monitoring has been shown to have high sensitivity and positive predictive value in 1 retrospective study,1 but this technique is not commonly used in North America. Recently, a noninvasive means of monitoring skeletal muscle perfusion with near-infrared spectroscopy (NIRS) became available, and its potential role in the diagnosis of ACS has been reported.16–21 It is hypothesized that monitoring skeletal muscle perfusion in a continuous fashion could greatly assist clinicians in making a timely and accurate diagnosis of ACS, so that early fasciotomy can be performed and unnecessary fasciotomy avoided. However, NIRS has not been widely studied in ACS, and combined monitoring using both NIRS and continuous pressure monitoring has not been reported, despite the potential value of these complementary technologies.
The Predicting Acute Compartment Syndrome Study was designed to develop a better understanding of the natural history of elevated compartment pressure and the utility of early and continuous monitoring of the physiologic status of the injured extremity in the timely diagnosis of ACS. In this study, the probability of ACS is determined retrospectively by an independent panel of experts based on the patient's initial presentation, clinical course, and known outcome at 6 months (including functional outcome and presence or absence of myoneural deficit). These probabilities are then modeled as a function of the clinical information obtained within the first 72 hours of injury to determine the predictive value of the data that would be available at the time of initial assessment. The immediate objective of the study is the development of a model that accurately predicts the likelihood of ACS based on data available to the clinician within the first 72 hours of injury, to include specific clinical findings supplemented by objective data obtained from state-of-the-art physiologic monitoring tools, specifically, muscle oxygenation measured using NIRS, and continuous monitoring of IMP and PP. The long-term objective of the research is to develop a tool that can aid clinicians in making a timely and accurate diagnosis of ACS.
METHODS: STUDY PROCEDURES AND PATIENT SELECTION
The study was conducted at 7 US trauma centers participating in the Major Extremity Research Consortium (METRC), which are listed in the Appendix 1.22 All study materials, including the protocol, consent, and recruitment materials, were developed in conjunction with the METRC Coordinating Center located at the Johns Hopkins Bloomberg School of Public Health and approved by the Department of Defense (DoD) Human Research Protection Office (study sponsor) and the local institutional review board (IRB) at each participating center. Local sites also obtained DoD Human Research Protection Office approval of local IRB documents and were thoroughly trained and then certified by the Coordinating Center before engaging in study activities to ensure proper execution of study procedures and data collection.
In this prospective study, no intervention was used to alter the treating physician's standard of care for the diagnosis or treatment of ACS. Physicians were blinded to continuous oxygenation and pressure data.
Patient Selection Criteria and Consent
The study inclusion and exclusion criteria are outlined in Table 1. Patients meeting the study inclusion criteria were approached for informed consent within 18 hours of injury. Legally authorized representatives were approached when patients were unable to provide informed consent. METRC has adopted a comprehensive informed consent process for all of its studies that involves the treating surgeon, the clinical site research coordinator, and material and resources for patients and family members to facilitate informed decision making about participation (see Figure, Supplemental Digital Content Figure 1, http://links.lww.com/BOT/A898 describes the METRC consenting procedures).
Study Procedures: In-Hospital Monitoring
After obtaining informed consent, 48 hours of monitoring was initiated. This monitoring included blood pressure, pain (using a visual analog scale23), neurovascular assessments including assessment of pulses, pain with passive stretch, muscle strength, and a sensory examination of the superficial peroneal, deep peroneal, sural, and posterior tibial nerves of the injured limb every 4 hours. Also obtained were daily creatinine phosphokinase values and continuous IMP monitoring and tissue oxygenation monitoring. The study injury was characterized using the AO/OTA fracture classification,24 Gustilo classification,25 and Tscherne soft tissue classification.26 Standard of care radiographic images were captured in the study database. All participants who underwent a surgical procedure to the study the limb during the initial monitoring interval were disconnected from monitors during the procedure. After the procedure, monitoring resumed for an additional 24 hours.
Continuous IMP was assessed using the Twin Star ECS (Bloomington, MN) Pressure Monitoring Fluid Collection (PMFC) catheter and ECS monitoring unit, which recorded IMP and the patient's blood pressure. The PMFC catheters were not connected to suction, and only the pressure monitoring functionality of the devices was employed in this study. Tissue oxygenation was monitored using the Nonin Medical Equanox 7600 (Plymouth, MN) with the 8003CA sensor. Both devices are cleared by the FDA for the study procedures; a nonsignificant risk determination was obtained to allow the PMFC catheters to be used for up to 48 continuous hours, as they are only FDA approved for 24 hours.
Continuous monitoring of both pressure and oxygenation was done only in the anterior compartment, according to the “sentinel compartment” principle, which relates to the fact that the anterior compartment typically is the earliest compartment involved in ACS and usually has the highest pressures.6,27 In addition, the deep posterior IMP was monitored with a pressure catheter because this compartment is also commonly involved and is the most difficult compartment to monitor clinically.
In this study, treating physicians were blinded to the NIRS and continuous measures of IMP and PP, and their diagnosis and treatment of ACS followed standard practice. Specifically, clinicians monitored the patient using standard clinical guidelines, and had the ability to use the IMP monitor to obtain up to 2 discreet measures of IMP if and when they determined that pressure measurement was clinically indicated as part of their standard management of patients at risk of ACS.
Additional data were collected on subjects who received a fasciotomy during the monitoring period, including prefasciotomy and postfasciotomy pressure (using a Stryker device), oxygenation (using the NIRS sensor), and clinical data, including appearance of tissue at the time of fasciotomy, and time to and method of wound closure. Photographs of the wound were also obtained. No additional compartment monitoring was conducted after the fasciotomy.
Study Procedures: Patient Follow-up
Participants were asked to return for follow-up study visits at 3 and 6 months after injury. At the 3-month study visit, the participants completed the Short Form Musculoskeletal Function Assessment (SMFA).28 At the 6-month visit, participants completed the SMFA again, and muscle strength was assessed using a handheld dynamometer (MicroFET2 MMT) and single leg heel rise tests.29 A clinical examination was conducted to assess the presence of myoneural deficit (based on sensory and motor examination of superficial peroneal, deep peroneal, sural, posterior tibial nerves), complications, wound healing, and fracture healing (defined as healed if no pain with weight-bearing, bridging callus on 3 cortices). Photographs were taken of the leg to include the surgical incisions and/or wound site, and follow-up radiographs were obtained. For patients who did not return to clinic, a combination of chart review and phone calls was used to assess gross myoneural function, including use of a device or orthotic to assist in ambulation, and pain.
The protocol underwent several modifications, most of which were administrative in nature. Most significantly, after 100 patients were enrolled, the protocol was amended to limit enrollment to participants at a higher suspected risk of ACS and to expand the window for enrollment from 12 to 18 hours postinjury. Specifically, patients with distal tibia fractures were excluded. These changes overcame an initial barrier to enrollment and allowed a focus on enrolling the individuals most likely to develop the outcome of interest.
METHODS: ASSIGNING PROBABILITY OF ACS
Once collected, data were cleaned and displayed on a custom online application designed for this study to present the data in a 2-stage process to allow reviewers to assess the likelihood that each participant had experienced a compartment syndrome based first on clinical data typically available (injury characteristics, radiographs, pain medications administered, neurovascular assessment data, surgical procedures, and myoneural and functional outcomes), and then with the addition of continuous oxygenation and pressure monitoring data (Fig. 1). Nine clinical experts, grouped in 3 panels of 3, were asked specifically to (1) estimate the likelihood that the person had compartment syndrome (0–100 scale, where 50 = completely unsure); (2) identify the 3 most important aspects contributing to that decision; (3) indicate whether or not they thought the patient had compromised muscle perfusion that would have benefited from a fasciotomy; (4) indicate whether they would have performed a fasciotomy if they had been the clinician responsible for the care the patient; and (5) indicate whether or not the pressure and/or oxygenation data influenced decision making, for the second round of scoring. Finally, reviewers were asked to identify whether or not data were missing that would have influenced decision making, and if so, what the impact would have been. Each panel reviewed 25% of the cases, and all 9 experts reviewed the remaining 25% of cases to allow for calibration of ratings across panels. When a panel was discordant in their assessment of likelihood of ACS (defined as a range of scores >30%), the group was brought together to review the case, and the case was rescored by all 9 experts. The goal of rescoring was not to force consensus but rather to allow reviewers to discuss aspects of the case that influenced decision making and prevent a single score from unduly influencing results.
METHODS: DATA MANAGEMENT, SAMPLE SIZE, AND ANALYSIS
Data were collected by site research coordinators and clinical investigators using paper case report forms designed specifically for this study and then entered into REDCap,30 the web-based data collection system used in all METRC studies. Monitoring data were downloaded from the devices and uploaded to REDCap (see Figure, Supplemental Digital Content Figure 2, http://links.lww.com/BOT/A899).
The sample size was determined using a simulation study. Various sample sizes, ranging from 100 to 1000, were considered. For each proposed sample size, oxygenation and pressure levels at 2 time points were generated for each patient, and, based on these levels, an average probability of ACS (as would be assigned by the expert panel) was generated. A regression based on these data was then built and evaluated for its predictive ability on a hypothetical external validation dataset of 2000 patients. There was an 8% improvement in predictive ability when increasing the sample size from 100 to 200, a 2% improvement between 200 and 300, and even smaller percentage improvements for larger sample sizes. With these results, the study was designed to enroll 200 patients.
The time series of pressure and oxygenation will be summarized based on both clinical and data-driven criteria. The outcome of the panelists' reviews will be modeled using generalized mixed effects models with linear31 and regression tree structures.32 These models will include random effects to account for the fact that each patient has multiple raters, and each rater is evaluating multiple patients. “Optimal” prediction models will be developed based on (1) only clinical variables that are available before making the decision to perform a fasciotomy and (2) these variables plus the summaries of pressure and oxygenation. The “value-added” of the models which include pressure and oxygenation data will be formally evaluated. The models will be developed using cross-validation techniques. Other prediction approaches (eg, random forests, bagging, and boosting32) will be investigated. Multiple imputation methods will be used to handle missing data.33–35
STUDY ENROLLMENT AND BASELINE DATA
In total, 191 participants from 7 centers were enrolled in this study (32% of those eligible); 48% of eligible patients were not enrolled because of administrative reasons, primarily related to lack of availability of equipment or staff to enroll a participant within the 18 hours window (Fig. 2). Final outcomes were obtained on 181 (95%) participants, 163 (85%) of which were completed in person. Participants were monitored for an average of 37.4 hours (SD 16 hours), and 24 (12.6%) received a fasciotomy. The average age was 38.1 years (SD 12.8). The study sample was 75% male, most of injuries were motor vehicle-related (28% occupant, 22% pedestrian or cyclist, and 22% motorcyclist). Nearly one half (49%) of participants overall had a closed fracture. Patient characteristics are summarized in Table 2.
This study has many strengths. The prospective collection of clinical signs (pain, motor strength, sensation) provides more reliable information than can be obtained from retrospective chart review. Furthermore, the continuous monitoring of both IMP and oxygenation using NIRS will provide new information for clinicians to consider regarding the utility of these tests in predicting the likelihood of ACS. Most importantly, the blinded collection of these continuous assessments affords the opportunity to examine the utility of these measures without influencing actual clinical decisions regarding whether or when to perform fasciotomy.
An additional strength of the study is that it addresses one of the most significant challenges in compartment syndrome research. Although there are proposed indicators for when to perform fasciotomy,1,5,27 there are no validated gold-standard definitions that reliably indicate when ACS is either present or an imminent diagnosis. Lacking such a standard, the performance of fasciotomy has been typically considered synonymous with the diagnosis of compartment syndrome. In this study, the predictive value of the various types of clinical data that are available after injury will not be determined until there has been an independent assessment by a panel of experts as to whether the patient actually had a compartment syndrome, based not only on the patient's initial presentation and clinical course, but also on the known outcome data for each patient.
Although the predictive value of clinical findings,8 discrete IMP measurements,7,9 and continuous pressure monitoring1 have been evaluated independently, no study has yet evaluated the reliability of NIRS for diagnosis of ACS, and no study has combined these assessments together into a single clinical decision rule. In addition, the consensus panels are directed to distinguish between cases where they believe there was compromised muscle perfusion that would have benefited from a fasciotomy versus cases where a “prophylactic” fasciotomy was conducted based on the risk of compartment syndrome developing in the future. This will allow some investigation of any remaining gaps in the assessment of compartment syndrome and provide some insight into clinical decision making.
One of the main weaknesses of this study is the lower-than-expected incidence of fasciotomy. Midway through the study, the study inclusion criteria were modified, with a goal of increasing the percentage of cases receiving fasciotomy (and by proxy, increasing the number of likely compartment syndrome cases) in the study. Although this increased the number of fasciotomies, overall fasciotomy rate remained low (12.6%).
Although the prediction model to be developed as part of this study will require validation in an independent sample of patients, it has the potential to define a diagnostic standard for the treatment of compartment syndrome, both in civilian practice and in the treatment of patients with military orthopedic injuries. The long-term goal of this research is to integrate the classic clinical signs and symptoms of ACS (which by themselves are of uncertain clinical significance) with the latest advances in monitoring and treatment of ACS: NIRS and continuous monitoring of IMP. The long-term goal of deriving a validated prediction rule tool for ACS will benefit patients and surgeons by reducing the incidence of both unnecessary fasciotomy and missed ACS. Furthermore, the insights gained by evaluating both the usefulness of various data points and the clinical decision-making process with regard to diagnosing ACS and determining the need for fasciotomy will inform further research that may in turn lead to refinements in the clinical care of these complex patients.
APPENDIX 1. CORPORATE AUTHORS
Participating Centers: Carolinas Medical Center: Michael J. Bosse, MD, Christine Churchill, MA, Joseph R. Hsu, MD, Rachel B. Seymour, PhD, Madhav A. Karunakar, MD, Stephen H Sims, MD; Hennepin County Medical Center: Andrew H. Schmidt, MD, Gudrun E Mirick, MD, Jerald R. Westberg, BA; San Antonio Military Medical Center: Daniel J. Stinner, MD; Denver Health Medical Center: David J. Hak, MD, MBA, FACS; University of Maryland R Adams Cowley Shock Trauma Center: Robert V. O'Toole, MD, Theodore Manson, MD, Timothy G. Costales, MD, Amanda C. Holmes, MS, Jason W. Nascone, MD, Andrew G. Dubina, MD; Vanderbilt Medical Center: William T. Obremskey, MD, MPH, MMHC, Eduardo J. Burgos, MD, A. Alex Jahangir, MD, MMHC, Hassan R. Mir, MD, MBA, FACS (now at University of South Florida Department of Orthopaedic Surgery), Rajesh R. Tummuru, MBBS, Manish K. Sethi, MD; Wake Forest Baptist Medical Center: Eben A. Carroll, MD, J. Brett Goodman, MBA, Jason J. Halvorson, MD, Martha B. Holden, AAS, AA, Anna N. Miller, MD, FACS (now at Washington University in St. Louis School of Medicine); Other Corporate Authors: Brown University Warren Alpert School of Medicine: Roman A. Hayda, MD (protocol committee member); METRC Coordinating Center at Johns Hopkins Bloomberg School of Public Health: Ellen J. MacKenzie, PhD, Lauren E. Allen, MA; Anthony R. Carlini, MS, Renan C. Castillo, PhD, Susan Collins, MSc, Katherine P. Frey, RN, MPH, Grace K. Ha, PhD, Daniel O. Scharfstein, ScD, Vadim Zipunnikov, PhD.
1. McQueen MM, Duckworth AD, Aitken SA, et al. The estimated sensitivity and specificity of compartment pressure monitoring for acute compartment syndrome. J Bone Joint Surg Am. 2013;95:673–677.
2. Allmon C, Greenwell P, Paryavi E, et al. Radiographic predictors of compartment syndrome occurring after tibial fracture. J Orthop Trauma. 2016;30:387–391.
3. Blair JA, Stoops TK, Doarn MC, et al. Infection and nonunion after fasciotomy for compartment syndrome associated with tibia fractures: a matched cohort comparison. J Orthop Trauma. 2016;30:392–396.
4. Crespo AM, Manoli A III, Konda SR, et al. Development of compartment syndrome negatively impacts length of stay and cost after tibia fracture. J Orthop Trauma. 2015;29:312–315.
5. Kosir R, Moore FA, Selby JH, et al. Acute lower extremity compartment syndrome (ALECS) screening protocol in critically ill trauma patients. J Trauma. 2007;63:268–275.
6. McQueen MM, Gaston P, Court-Brown CM. Acute compartment syndrome. Who is at risk? J Bone Joint Surg Br. 2000;82:200–203.
7. Prayson MJ, Chen JL, Hampers D, et al. Baseline compartment pressure measurements in isolated lower extremity fractures without clinical compartment syndrome. J Trauma. 2006;60:1037–1040.
8. Ulmer T. The clinical diagnosis of compartment syndrome of the lower leg: are clinical findings predictive of the disorder? J Orthop Trauma. 2002;16:572–577.
9. Whitney A, O'Toole RV, Hui E, et al. Do one-time intracompartmental pressure measurements have a high false-positive rate in diagnosing compartment syndrome? J Trauma Acute Care Surg. 2014;76:479–483.
10. O'Toole RV, Whitney A, Merchant N, et al. Variation in diagnosis of compartment syndrome by surgeons treating tibial shaft fractures. J Trauma. 2009;67:735–741.
11. Bhattacharyya T, Vrahas MS. The medical-legal aspects of compartment syndrome. J Bone Joint Surg Am. 2004;86-A:864–868.
12. Olson SA, Glasgow RR. Acute compartment syndrome in lower extremity musculoskeletal trauma. J Am Acad Orthop Surg. 2005;13:436–444.
13. Heemskerk J, Kitslaar P. Acute compartment syndrome of the lower leg: retrospective study on prevalence, technique, and outcome of fasciotomies. World J Surg. 2003;27:744–747.
14. Ritenour AE, Dorlac WC, Fang R, et al. Complications after fasciotomy revision and delayed compartment release in combat patients. J Trauma. 2008;64(suppl 2):S153–S161; discussion S161–2.
15. Finkelstein JA, Hunter GA, Hu RW. Lower limb compartment syndrome: course after delayed fasciotomy. J Trauma. 1996;40:342–344.
16. Arbabi S, Brundage SI, Gentilello LM. Near-infrared spectroscopy: a potential method for continuous, transcutaneous monitoring for compartmental syndrome in critically injured patients. J Trauma. 1999;47:829–833.
17. Garr JL, Gentilello LM, Cole PA, et al. Monitoring for compartmental syndrome using near-infrared spectroscopy: a noninvasive, continuous, transcutaneous monitoring technique. J Trauma. 1999;46:613–616; discussion 617–8.
18. Gentilello LM, Sanzone A, Wang L, et al. Near-infrared spectroscopy versus compartment pressure for the diagnosis of lower extremity compartmental syndrome using electromyography-determined measurements of neuromuscular function. J Trauma. 2001;51:1–8; discussion 8–9.
19. Giannotti G, Cohn SM, Brown M, et al. Utility of near-infrared spectroscopy in the diagnosis of lower extremity compartment syndrome. J Trauma. 2000;48:396–399; discussion 399–401.
20. Shuler MS, Reisman WM, Whitesides TE Jr, et al. Near-infrared spectroscopy in lower extremity trauma. J Bone Joint Surg Am. 2009;91:1360–1368.
21. Shuler MS, Reisman WM, Kinsey TL, et al. Correlation between muscle oxygenation and compartment pressures in acute compartment syndrome of the leg. J Bone Joint Surg Am. 2010;92:863–870.
22. Major Extremity Trauma Research Consortium (METRC). Building a clinical research network in trauma orthopaedics: the major extremity trauma research consortium (METRC). J Orthop Trauma. 2016;30:353–361.
23. Knop C, Oeser M, Bastian L, et al. Development and validation of the visual analogue scale (VAS) spine score [in German]. Unfallchirurg. 2001;104:488–497.
24. Marsh JL, Slongo TF, Agel J, et al. Fracture and dislocation classification compendium—2007: orthopaedic trauma association classification, database and outcomes committee. J Orthop Trauma. 2007;21(suppl 10):S1–S133.
25. Gustilo RB, Anderson JT. Prevention of infection in the treatment of one thousand and twenty-five open fractures of long bones: retrospective and prospective analyses. J Bone Joint Surg Am. 1976;58:453–458.
26. Tscherne H, Oestern HJ. A new classification of soft-tissue damage in open and closed fractures (author's transl) [in German]. Unfallheilkunde. 1982;85:111–115.
27. McQueen MM, Christie J, Court-Brown CM. Acute compartment syndrome in tibial diaphyseal fractures. J Bone Joint Surg Br. 1996;78:95–98.
28. Swiontkowski MF, Engelberg R, Martin DP, et al. Short musculoskeletal function assessment questionnaire: validity, reliability, and responsiveness. J Bone Joint Surg Am. 1999;81:1245–1260.
29. Möller M, Lind K, Styf J, et al. The reliability of isokinetic testing of the ankle joint and a heel-raise test for endurance. Knee Surg Sports Traumatol Arthrosc. 2005;13:60–71.
30. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42:377–381.
31. McCullagh P, Nelder JA, eds Generalized linear models. Monographs on Statistics and Applied Probability. No 37. London, United Kingdom: Chapman and Hall; 1989.
32. Friedman J, Hastie T, Tibshirani R, eds The elements of statistical learning. Springer Series in Statistics. No 1. Berline, Germany: Springer; 2001.
33. Harel O. Inferences on missing information under multiple imputation and two-stage multiple imputation. Stat Methodol. 2007;4:75–89.
34. Rue T, Thompson HJ, Rivara FP, et al. Managing the common problem of missing data in trauma studies. J Nurs Scholarsh. 2008;40:373–378.
35. Schafer JL. Multiple imputation: a primer. Stat Methods Med Res. 1999;8:3–15.