Continuous Near-Infrared Spectroscopy Demonstrates Limitations in Monitoring the Development of Acute Compartment Syndrome in Patients with Leg Injuries

This article was updated on October 29, 2018, because of a previous error. On page 1645, in the group authorship footnote listing the members of the Major Extremity Trauma Research Consortium (METRC), the name “Anna N. Miller” was not included in the list of members. The list now reads “Major Extremity Trauma Research Consortium (METRC)Christine Churchill, Joseph R. Hsu, Rachel B. Seymour, Stephen H. Sims, A. Alex Jahangir, Robert H. Boyce, Manish K. Sethi, Andres Rodriguez-Buitrago, Vamshi Gajari, Jason W. Nascone, Marcus F. Sciadini, Theodore Manson, Timothy G. Costales, Merryjessica Fuerst, W. Andrew Eglseder, Christopher LeBrun, Andrew N. Pollak, J. Brett Goodman, Jason J. Halvorson, Martha B. Holden, Anna N. Miller, Jerald R. Westberg, Dennis Mann, and Susan Collins”.An erratum has been publishedJ Bone Joint Surg Am. 2018 Dec 5;100(23):e151. BACKGROUND:We recorded measurements of muscle perfusion using near-infrared spectroscopy (NIRS) and intramuscular pressure (IMP) in a study designed to develop a decision rule for predicting acute compartment syndrome (ACS). The purpose of this study was to report our experience measuring NIRS data in the context of this broader investigation and to explore factors related to variations in data capture. METHODS:One hundred and eighty-five patients with lower-leg injuries had data consisting of continuous NIRS measurement of the O2 saturation in the anterior compartment of the injured limb and the contralateral (control) limb, and continuous IMP recording in the anterior and deep posterior compartments of the injured leg as part of their participation in an institutional review board-approved multicenter trial. All monitoring was done for a prescribed period of time. For both types of data, the percentage of valid data capture was defined as the ratio of the minutes of observed data points within a physiological range to the total minutes of expected data points. Clinically useful NIRS data required simultaneous data from the injured and control limbs to calculate the ratio. Statistical tests were used to compare the 2 methods as well as factors associated with the percent of valid NIRS data capture. RESULTS:For the original cohort, clinically useful NIRS data were available a median of 9.1% of the expected time, while IMP data were captured a median of 87.6% of the expected time (p < 0.001). Excluding 46 patients who had erroneous NIRS data recorded, the median percentage was 31.6% for NIRS compared with 87.4% for