Fig. 1 The operational definitions used to categorize patients.
Courtesy of Jonathan G. Schoenecker, MD


Published 8/1/2014
Peter Pollack

Screening for Musculoskeletal Infection in the ED

Applying principles learned from damage-control orthopaedics to infection

"With all the research that’s been done during the last 20 years, we’ve become facile at identifying infected patients from those without infection,” explained Jonathan G. Schoenecker, MD. “However, we have observed that all infections are not created equal.”

Dr. Schoenecker presented findings from his study, “Screening Inflammation and Coagulation Activity Allows for Rapid Risk Stratification in Children Presenting with Symptoms of Musculoskeletal Infection,” at the annual meeting of the Pediatric Orthopaedic Society of North America.

“Some infections just require a little débridement,” he continued, “while others end up in intensive care units (ICUs). Current diagnostic algorithms are not adequate at delineating these two groups. To provide the best care for these patients, we need to find methods of determining the severity of infections quickly and early. This would enable the development of specific care pathways designed to save lives and limbs and to provide the most efficient and cost-effective medicine possible.

“We set out to develop an algorithm designed to also segregate patients with minor infections from those with severe infections,” he continued. “We first had to develop a classification system for children with musculoskeletal infection that was inclusive of all musculoskeletal tissue—muscles, joints, and bones. We then tested the hypothesis that we could find predictors in the emergency department (ED) that would enable us to categorize patients with severe infections early in the disease process and treat them accordingly.”

Classifying infection
The first challenge Dr. Schoenecker and his colleagues faced was developing a way to classify the severity of a musculoskeletal infection.

“At first, we thought we would use permanent joint disability as a sorting factor,” he said, “but that’s actually a fairly rare event.”

The team instead developed a theory that allowed them to create operational definitions for infection severity.

“In this theory, we consider musculoskeletal infection as a tissue injury,” he said. “Based on what we’ve learned from damage-control orthopaedics, the area that is infected is less important than how much of the musculoskeletal tissue is actually injured. In this model, the injury causes an acute phase response that triggers interleukin-6 (IL-6) to modify around 1,000 hepatic genes that provide hemostasis and antimicrobials to fight that infection. The difference between infection and a fracture, however, is that infection is a persistent tissue injury, which increases the acute phase response far more than a fracture. “

Dr. Schoenecker explained that, as infection spreads, so does tissue injury, resulting in a corresponding increase in the acute phase response. Further, the tissue injury continues until it is resolved by the host immune system, antibiotics, or potentially surgery.

The retrospective study of more than 200 pediatric patients evaluated for musculoskeletal infection categorized patients using operational definitions designed to quantify the amount of tissue injury caused by infection (Fig. 1).

“The aseptic group was defined by negative local and blood cultures,” he said. “The ‘local’ infection group was defined by infection of one anatomic compartment and no more than one positive blood culture. Patients with a ‘disseminated’ infection had multifocal disease, either infection of multiple anatomic compartments, thromboembolic disease, and/or two or more positive blood cultures.”

After classifying each patient based on level of infection, Dr. Schoenecker and his team reviewed inpatient hospital outcomes and found that patients in the disseminated infection cohort had significantly longer ICU times, hospital stays, and days on antibiotics. They also had more surgeries and increased hospital costs.

Delineating infection levels
After determining that their operational definitions were valid, the research team reviewed the patient records to determine whether they could have correctly classified each patient in the ED. As expected, and similar to what has been observed in damage-control orthopaedics, they found that monitoring changes in the acute phase response correlated well with the amount tissue injured by infection.

“Based on our patient population, we found that a C-reactive protein (CRP) of 20 was an excellent delineator of aseptic versus infected, while a CRP of 90 was able to separate local and disseminated infections when the patient was initially seen. Using our model, we noted that the more overwhelming tissue injury leads to an increase in CRP, which is a surrogate for IL-6, the key regulator of the acute phase response.”

Furthermore, applying the principle that an extreme and persistent acute phase response causes a massive activation and consumption of clotting factors, the group determined measured D-dimer levels, a surrogate for coagulation activation. They found that D-dimer levels of greater than 1.2 μg/mL also accurately stratified local and disseminated infections. However, temperature, white blood cell count, and erythrocyte sedimentation rate were not predictive of local versus disseminated infection.

“Our next step,” he said, “will be to conduct a multicenter study to determine if these predictors hold true in a prospective manner. We can then hopefully develop different treatment regimens that provide the best care for each category of infection.”

Dr. Schoenecker credits the group effort by Megan E. Mignemi, MD; Tommy An; Michael A. Benvenuti, Derick J. Williams, MPH, MD; and the Vanderbilt Pediatric Orthopaedic division for the thoroughness and success of this study. The results and future of this work will be highlighted at the 2015 combined AAOS/ORS symposium in Las Vegas.

Disclosure information: Dr. Schoenecker—ISIS pharmaceuticals.

Peter Pollack is the electronic content specialist for AAOS Now. He can be reached at

Bottom Line

  • Based on damage-control orthopaedics, the area that is infected is less important than how much musculoskeletal tissue is actually injured.
  • Patients with disseminated infection had longer ICU times, hospital stays, and days on antibiotics, along with more surgeries and increased hospital costs.
  • The level of acute phase response factors CRP and D-dimer appear to correlate to degree of infection at the time the patient is first seen in the ED.