Introduction
The most frequent complication of open fractures, with a reported incidence of 3 to 40%, is infection1. The anatomy of the tibia with limited soft tissue coverage and marginal blood supply that can be easily disrupted from soft tissue injury predisposes open fractures of the tibia to high rates of infection (Newman Injury). Open tibia fractures exhibit infection rates ranging 10-20 times higher than other locations 2. Infection of an open fracture can be devastating, potentially leading to osteomyelitis, systemic infection, amputation, or even death3. Therefore, an open fracture is one of a few orthopedic injuries that requires timely treatment.
Current open fracture treatment protocols aim to reduce the risk of infection through emergent intravenous (IV) antibiotics and surgical debridement 4. Over time, an unwritten “six-hour rule” has developed within the orthopedic community; it is widely advocated that initiating treatment within six hours of injury reduces the rate of morbidity and mortality5-8. While the literature agrees that urgent care is essential, there is little scientific evidence to support this “six-hour rule” 1, 4, 5, 8-12.
Soddo Christian Hospital (SCH) in Wolaita, Soddo, Ethiopia provides a unique opportunity to examine the influence of prolonged time to treatment on infection in open tibia fractures. SCH is a tertiary referral center in Southern Ethiopia with a strong orthopedic service. Rugged terrain and limited means of transportation often require patients to travel long distances, sometimes requiring many hours or even days, to receive care. For example, one study examining tibia fractures at hospitals in three developing countries, including SCH, found an average time from injury to surgery of 4.1 days 13. In addition, SCH provides a controlled environment; relatively few surgeons operate at SCH and the majority of tibia fractures are treated with Surgical Implant Generation Network (SIGN) intramedullary nails. The purpose of this study is to retrospectively examine the effect of delayed treatment on infection rates in open tibia fractures on a scale that has not previously been examined.
Materials and Methods
Soddo Christian Hospital (SCH) is a mission hospital that operates in Soddo, Ethiopia. As one of the only surgical hospitals in southern Ethiopia, it serves as a major referral center for trauma and orthopedic surgery. SCH is also one of 287 hospitals in 50 developing countries that utilize the Surgical Implant Generation Network (SIGN) intramedullary nail for fracture fixation14.
All patients 18 years or older, who presented to SCH from January 2006 to June 2015 with a unilateral open tibia fracture (AO/OTA 41A, 42A-C, 43A) and were treated with the SIGN intramedullary nail were eligible for this study15. Exclusion criteria included patients with bilateral tibia fractures, a concomitant femur fracture, patients that presented greater than 120 days from the time of their injury, nonunions, and pathologic fractures. Although no formal treatment protocol existed, patients were treated based upon community standards. Patients were emergently assessed for life-threatening injury by the emergency and surgical teams and stabilized as needed. IV antibiotic prophylaxis (cloxacillin and gentamycin) was initiated and continued at the discretion of the treatment team, generally 24-72 hours. Following initial stabilization, patients were urgently irrigated and surgically debrided. Repeat surgical debridement and irrigation was performed as necessary. Patients were graded according to the Gustilo and Anderson classification system by the orthopedic team at SCH 16. All fractures were stabilized using the SIGN intramedullary nail. The method of wound closure varied depending on soft tissue injury and was determined by the surgical team. Closure was achieved by one of the following methods: primary closure, delayed primary closure, secondary closure, skin grafting, or flap closure. Patients were encouraged to return for follow up at one month and subsequent regular intervals. All surgical and follow up data was prospectively collected and recorded in the SIGN database.
Any patient that received antibiotics for greater than 72 hours were considered to have developed an infection. Deep infection was diagnosed clinically by the attending surgeon considering physical exam (fever, erythema, edema, pain), laboratory analysis (leukocytosis), or x-ray imaging (evidence of osteomyelitis or hardware loosening) at any follow-up visit greater than four weeks after treatment. Infectious complications were recorded prospectively in the SIGN database but immediate postoperative infection was determined by retrospective analysis.
All data was retrospectively collected by the authors from SCH’s SIGN database and consolidated in a secure REDCap database 17. Time to treatment was determined by the interval from the time of injury to first antibiotic administration. Data and statistical analysis was performed using Microsoft Excel (Microsoft Office for Mac, Version 15.32) and SPSS (IBM, Version 23). A fisher’s exact test was used to evaluate the difference in infection rates of open fractures treated within 24 hours and those treated at 24 hours or greater. Further analysis using a 2 test for trend was employed to evaluate the association between Gustilo type and infection. Lastly, an unpaired t test was used to assess the difference in time to treatment of infected and noninfected fractures.
Results
One hundred and seventy-two open tibia fractures were retrospectively reviewed. The average patient age was 32 years (range, 18-70). The population consisted of 132 (77%) male and 40 (23%) female patients. There were 44 (26%) type I fractures, 39 (23%) type II fractures, 44 (26%) type IIIa fractures, 44 (26%) type IIIb fractures, and 1 (1%) type IIIc fracture. (Table 1)
There were 8 (5%) AO/OTA 41-A fractures, 93 (54%) AO/OTA 42-A-C fractures, and 52 (30%) AO/OTA 43-A fractures. Three fractures occurred in the middle one-third and extended to the distal one-third (42-C?). Twelve fractures were segmental, two of which occurred in the distal one-third. The location was not recorded for four fractures (Table 1). The cause of injury was not documented.
One hundred and thirteen patients (66%), 86 (76%) males and 27 (24%) females, had sufficient follow-up. The follow up population consisted of 25 (22%) type I fractures, 30 (27%) type II fractures, 30 (27%) type IIIa fractures, 28 (25%) type IIIb fractures, and zero (0%) type IIIc fractures. The average time from surgery to the first follow-up evaluation was 12.1 weeks. The average number of follow-up visits was 1.3 with an average total follow-up duration of 18 weeks (range, 4-155 weeks).
In total, nineteen (11%) fractures were determined to be infected. Thirteen (68%) infections occurred in the immediate post-injury hospital course and six (32%) patients presented with a deep infection at follow-up. Four (67%) of the deep infections presented at the first follow-up visit (mean, 10 weeks; range, 6-17 weeks) and two (33%) fractures were found to be infected at the second follow-up visit (mean, 45 weeks; range, 33-58 weeks). Infection occurred in one (2%) of type I fractures, six (15%) of type II fractures, nine (21%) type IIIa fractures, and three (7%) type IIIb fractures. The one type IIIc fracture did not become infected. There was not a significant correlation between increasing Gustilo type and infection (P = 0.402).
One hundred and forty-four (84%) patients received initial treatment within 24 hours of injury (mean, 7.0 hours) while 28 (16%) patients did not receive initial treatment until after twenty-four hours (mean, 41.1 hours; 29.3 without outlier of 360). Fifteen (10%) of the patients treated within 24 hours and four (14%) of the patients treated after 24 hours from the time of injury developed an infection. The rate of infection was not significantly different (P = 0.518) between the two groups. The average time to treatment of the infected and noninfected fractures was 15.6 and 13.4 hours, respectively. Unpaired t-test analysis showed that there was not a significant difference in the time to treatment of these groups (P = 0.700).
Discussion
Infection of an open tibia fracture the most important potential complication and can lead to severe consequences. Many articles have previously examined factors that influence the development of infection in open tibia fractures and how orthopedic trauma surgeons can prevent it. Treatment protocols advocate for expedient surgical debridement and IV antibiotics to reduce the risk of infection. However, over time, a “six-hour rule” seems to have evolved based mainly on historical tradition rather than scientific evidence 1, 4, 5, 8-12.
To our knowledge, no studies have examined delays in treatment of the magnitude studied in this paper. This is most likely due to the ethical implications of withholding treatment for the benefit of assessing the point at which infection becomes a greater risk. Our study utilized the naturally delaying factors of the Ethiopian geography and limited access to care to help further elucidate the relationship between treatment delay and infection in open fractures. Our results show that treatment delay beyond twenty-four hours does not lead to a significantly increased risk of infection. Current literature further supports this argument.
Several studies have directly challenged the idea that reduced time to treatment leads to fewer infections. A retrospective study of 191 open tibia fractures showed no significant increase in infection with respect to time from injury to initial operative management 1. Similarly, Bednar and Parikh showed no significant increase in infection rate with late irrigation and debridement 5. More recently, a prospective 5-year study examining infection in open long bone fractures did not find statistical evidence of an increased infection rate in fractures debrided greater than six hours from the time of injury 10. Conversely, one investigation did show a significant increase in the occurrence of infection debrided more than five hours after the initial injury 6. However, it is important to note that this study only examine patients with the more severe Gustilo type II and III tibia fractures, which may have led to biased results7
Although our results did not show a statistically significant association between increased fracture severity and increased rate of infection, our results do show a trend; a greater proportion of infections occurred in the more severe, higher Gustilo types. In fact, all but one infection occurred in type II and III fractures. This trend is further supported in the literature. A recent study created a multivariate model to evaluate the influence of several factors on infection rates 12. Gustilo type was found to be the greatest predictor of nonunion and infection, suggesting that management of open tibia fractures should be guided by Gustilo type rather than time from injury 12. The limitations of our study may have hidden the true association between Gustilo type and infection.
Our study is not without limitations. One these is loss to follow up, however, our rates of follow up (66%) are similar to those of other SIGN database studies in developing countries 18, 19. Furthermore, the retrospective design limited the amount and type of data that could be collected. For example, infection during the immediate post-operative hospital stay was not documented in the SIGN database necessitating that we use duration of antibiotic coverage as a proxy for the presence of infection. In addition, a community standard protocol supervised by two attending orthopedic surgeons was used to guide treatment but no formal protocol for surgical debridement or wound management was employed. It is important to note that all patients were treated similarly with expedient debridement, antibiotic prophylaxis, and fixation with the SIGN intramedullary nail creating a relatively controlled study environment. Moreover, it was difficult to control for confounding variables such as previous medical conditions, trauma history, tobacco use, and mechanism of injury. Despite these limitations, our study does provide valuable information on the effects of treatment delay beyond 24 hours. Our results suggest fracture type rather than time to treatment is a more important predictor of the risk of infection.
Conclusion
Ethiopia, with its rugged terrain and difficult transportation, provided a unique environment that allowed analysis of treatment delays that, for ethical reasons, would not be possible in a developed nation. The findings of the study suggest that significant delays in time to treatment of open tibia fractures is not associated with an increased risk of post-operative infection. While factors may limit the significance of this study, our results are consistent with current literature. The trend seen with our data further supports the idea that fracture Gustilo type may be the best predictor of infection risk. Further research is needed to determine how this can best guide management of open tibia fractures, both in the developed and developing worlds. It is our belief that expedient treatment should be employed in all cases but we contend that growing evidence does not support a “6-hour rule”.