Open Fractures: Still Risky After All These Years

Natoli et al. report infection and delayed union/nonunion rates of 17% and 9%, respectively, across a wide range of 2,381 open tibial fractures, with use of data combined from 3 multicenter prospective trials recently published in high-quality journals. This study was unique because of the high quality of the methodology and the granularity of the data. The volume of patients is so great that it is possible to identify infection and delayed union/nonunion rates as a function of the Gustilo-Anderson Open Fracture Classification and many of the OTA Open Fracture Classification1 categories at the bone-segment level. This article featured a novel methodology. The original studies were designed to investigate various forms of skin preparation and irrigation in the treatment of open fractures. This article reexamined pooled data from these similar studies to investigate a different question, namely: what are the rates of infection and delayed union/nonunion in a large series of patients with open fractures treated with current methods? The authors reference another recent publication assessing the rates of infection and delayed union/nonunion in 11,000 open tibial fractures over the past 4 decades. That study showed a 12% overall infection rate and 9% nonunion rate in the most recent decade, which was minimally changed from the 14% infection rate and 13% nonunion rate in the 1980s2. I agree with the authors’ conclusion that this new high-quality evidence supports the notion that infection and nonunion rates for open fractures remain high, with minimal change from historical rates. This finding is quite disappointing considering the amount of money and effort that has been expended over the past 40 years in an attempt to improve outcomes for patients with open fractures. It is now commonly accepted that early administration of antibiotics is one of the most important—if not the most important—interventions to minimize the risk of subsequent infection. Early antibiotic administration was practiced in all 3 of these studies as a standard of care, in contrast to open fracture protocols pre-2000. The conclusions of the article by Natoli et al. call into question the true efficacy of early antibiotics in reducing infection rates. It is also noted in both this study and the referenced study that the current standard of initial open fracture fixation is primarily open reduction and internal fixation (ORIF), whereas external fixation was more commonly utilized prior to 2000. The authors concluded that these data support the notion that ORIF of open fractures is not associated with an increased rate of infection. Another theoretical possibility is that the rate of infection actually does increase with ORIF as compared with external fixation, but that the overall rate of infection is the same because other factors (such as early antibiotics) have mitigated that increase in the rate of infection. However, the study by Natoli et al. was not specifically designed to determine the individual effects of any specific treatment, only the overall rates with currently accepted treatment concepts. This study was also not designed to determine the overall rate of open fractures. Neither the population at risk nor the percentage of patients with open fractures successfully recruited to participate in the prospective studies were measured. However, the pattern of severity of open fractures encountered was reported and is very similar to historical controls. This finding suggests that society’s investment in better roads, safer cars, and safer workplaces has not necessarily reduced the severity pattern of open fractures. The clinical impact of a post-fracture infection may have changed over the decades. In the past, infections were considered a disastrous complication with generally poor outcomes and were something to try to avoid at almost all costs. Chronic osteomyelitis was considered a contraindication to arthroplasty, for example. Today, post-fracture infections are considered more of an intermediate-severity complication that is generally amenable to successful treatment and salvage. It may be that we have rightly or wrongly become more accepting of post-fracture infections and are willing to accept the occurrence of some infections for the convenience of internal over external fixation. It was thought that early internal fixation of open fractures would be associated with a lower nonunion rate than external fixation, but that was not found in this study (reported as historical controls) or the referenced study. This finding suggests that the benefits of early internal rather than external fixation may be more convenience-based than outcome-based. Granularity is the ability to make clinically relevant conclusions regarding particular subsets of the data, such as individual Gustilo classes, and specific fracture types, such as tibial shaft fractures. These data show an increasing risk of infection with Gustilo fracture type, from 5% in Type 1 to 26% in Type 3B. Similar results were found for nonunion, with rates of 3% in Type 1 to 17% in Type 3C. Gustilo Type 3C is characterized by open fractures with arterial injuries requiring repair and were noted by Gustilo and Anderson to have a particularly high rate of amputation. These data are not particularly helpful in determining the current rate of amputation in Type 3C because the studies excluded patients with early amputation and only included failed attempts at limb salvage. The article by Natoli et al. and each of the referenced studies utilized the Gustilo-Anderson Open Fracture Classification to help direct treatment and report results, and none of them utilized the newer and perhaps better OTA Open Fracture Classification. Many shortcomings have been identified regarding the use of the Gustilo Classification since its publication 50 years ago. The Gustilo Classification is primarily based on the length of the skin wound. The OTA Open Fracture Classification incorporates other important criteria, such as contamination, arterial injury not requiring repair, bone loss, and other criteria that are considered prognostically important. We do not know whether the use of the OTA Open Fracture Classification to direct treatment might improve outcomes because it has not achieved widespread acceptance or implementation.