CORR Insights®: Ninety-Day Follow-up Is Inadequate for Diagnosis of Fracture-related Infections in Patients with Open Fractures

Where Are We Now?

Fracture-related infections are associated with pain, morbidity, additional unplanned surgical procedures, and, sometimes, death [3]. They also are associated with prolonged hospitalization and substantially increased treatment costs [9], which often are borne by healthcare systems as many patients who experience serious trauma in the United States do not have healthcare insurance [2]. Numerous recently published and ongoing multicenter trials have focused on the role of surgical interventions potentially minimizing the risk of fracture-related infections, such as local antibiotic application, oxygen supplementation, and contemporary wound management [4, 10, 11]. Nonetheless, fracture-related infections remain an unsolved clinical problem.

One important knowledge gap has been the duration of surveillance needed to ensure that most fracture-related infections are not missed. As Dr. Zalavras and colleagues [12] correctly emphasized in this month’s Clinical Orthopaedics and Related Research®, previous Centers for Disease Control and Prevention (CDC) guidelines recommended a surveillance period of 12 months [5], other experts may request longer surveillance periods [8], whereas the most recent CDC guidelines suggest a 90-day surveillance period to monitor for implant-related infections [1]. Current author instructions for CORR® request 12 months of follow-up data for clinical research papers reporting on infections, which is in line with recent recommendations published by an international expert panel [8]. Establishing guidelines for diagnosing and recording infections remains crucial for high-quality patient care and for standardizing research methods. Therefore, guidelines need to be based on best-available clinical evidence. Moreover, it remains the task of the orthopaedic community to critically evaluate established guidelines.

In their manuscript, Zalavras et al. [12] challenge current CDC guidelines; in this CORR paper, the authors analyzed the timing of onset of fracture-related infections in open long-bone fractures among 530 patients treated between 2003 and 2017. The authors found that 90 days of follow-up would have captured only 64% of fracture-related infections, whereas 12 months of follow-up would have captured 89% of fracture-related infections. The authors conclude that in patients with open long-bone fractures, follow-up of 90 days is inadequate for surveillance of fracture-related infections, especially for research purposes, and that 12 months of follow-up would be preferable. These findings also potentially carry important clinical implications as surgeons may elect to tailor their protocols for postoperative follow-up visits accordingly and follow patients routinely for longer time periods.

Where Do We Need To Go?

While Dr. Zalavras and colleagues [12] have made a discovery that is directly applicable to clinical research, a number of controversies remain. One obvious but important one is the very definition of fracture-related infection. Zalavras and colleagues [12] appropriately lean on specific confirmatory criteria including the presence of (1) positive cultures, (2) sinus tract, or (3) purulence, as suggested by an international expert panel [7]. But numerous other definitions and diagnostic criteria are in use; a recently published systematic review demonstrated that only two percent of randomized controlled trials cited a validated definition to describe infection after fracture fixation [6], which emphasizes the need for standardization.

In addition, surgical infections are often categorized as superficial or deep. The CDC has offered some refinement on this, by dividing infections into three categories instead of two: superficial incisional, deep incisional, and organ/space infections [1]. However, these criteria are difficult to apply for fracture-related infections, particularly in anatomic areas where relatively superficial implant placement is common, such as ankle fractures. Here again, it will be the responsibility of the orthopaedic community to standardize the definitions for categorizing the type of infection.

Future studies will need to give the work of Zalavras et al. [12] a bit more context; for example, this study included only open fractures. What about other types of injuries, such as closed fractures undergoing surgery? There also are cost-benefit questions; it seems obvious that even longer follow-up will capture more infections. Is 1 year—89% capture—a high enough percentage? It probably depends on the specific type of injury and infection and the goals of the study in question. Future studies will need to help us decide how many of these infections we can “afford to miss” by not tracking patients out to a year or beyond.

Finally, we need to recognize how challenging it is for orthopaedic trauma surgeons to collect long-term follow-up data in clinical practice, particularly among practices that deliver care to patients who are indigent [13]. To this point, much of what we have learned about trauma care has come from busy, practicing trauma surgeons; we need to find ways to make the most out of “real-world data,” even when patients are lost to follow-up.

How Do We Get There?

Going forward, it will be the responsibility of the orthopaedic community to establish evidence-based guidelines for surveillance periods regarding fracture-related infections. Establishing widely accepted guidelines will assist investigators in the development of study protocols, it will standardize study results, and it will allow us to compare the results of different studies.

Future studies on fracture-related infection rates need to apply standardized definitions for the diagnosis of fracture-related infections and for how to categorize fracture-related infections. But since there is little consensus on this, and as the CDC criteria are rather broad and less geared toward fracture-related infections, I suggest that the American Academy of Orthopaedic Surgeons and orthopaedic specialty societies, such as the Orthopaedic Trauma Association and the Musculoskeletal Infection Society, provide joint position statements that create definitions that are applicable to fracture-related infection. Similarly, the leading orthopaedic journals should implement a shared set of author guidelines for reporting about fracture-related infections.

Zalavras et al. [12] studied only open fractures, and future studies need to address these same questions in closed fractures undergoing surgical fixation. One key challenge when studying infection rates in closed fractures will be that the infection rates are typically much lower than in open fractures. Consequently, larger datasets will be needed to obtain a sufficient number of events. To get there, data from multicenter studies and multi-institutional patient registries may be required. However, analyzing a large quantity of data does not always result in high-quality research. Trauma registries sometimes are limited by low granularity of information, erroneous data entry, missing data, and short episodes of care.

Despite the advantages of multicenter studies with standardized treatment protocols and multi-institutional registries, we should not lose sight of the benefits of single-institution and even single-surgeon series. These can provide valuable information about results of institution-specific protocols, detailed experiences of individual surgeons, and local patient populations facing special challenges, perhaps even challenges that generalize to other centers. I believe multicenter studies and data from registries complement, but cannot replace, individual and institutional experiences.