Abstract
Introduction Open hand fractures are anecdotally reported to have lower infection rates than open long bone fractures. Although a 3-hour rule for antibiotic administration and a 6-hour rule for operative debridement have historically been upheld as ideal management for open fractures, other factors may be more influential in the development of infection. The purpose of this study was to investigate factors associated with open hand fracture infections.
Materials and Methods We retrospectively reviewed 67 patients with 107 open hand fractures between 2012 and 2017. Time from injury to antibiotic administration and operative debridement, modified Gustilo–Anderson classification, and patient characteristics including age, smoking status, and presence of chronic disease were examined for each patient. Outcome parameters were the development of infection and fracture union.
Results The overall rate of infection was 9% (6 of 67 patients). No type 1 or type 2 fractures developed infection in contrast to 12.2% of type 3 fractures. Patients who received antibiotics in less than 3 hours and underwent debridement in less than 6 hours did not have lower infection or nonunion rates than those who did not. The association between the modified Gustilo–Anderson classification and the development of infection or nonunion was statistically significant.
Conclusion Factors including time to antibiotics, time to operative debridement, smoking status, and chronic disease comorbidities were not predictive of either infection or nonunion in open hand fractures. Fracture type as defined by a modified Gustilo–Anderson classification was the factor most strongly related to the development of infection or nonunion in these fractures.
Keywords: open fractures, hand fractures, trauma, infection, non-union
Introduction
Open fractures are difficult to treat and can be fraught with complications and devastating outcomes. One of the most serious complications is the development of infection, which impedes wound healing and, if untreated, can contribute to multiple complications including nonunion. 1 The majority of the literature on open fractures has focused on open long bone fractures. There is a lack of consensus on the appropriate classification and management of open hand fractures. Over the past several decades, several groups have either modified the Gustilo–Anderson open fracture classification system to the hand or have developed their own distinct classification system. For example, Swanson et al proposed a classification system based on fracture contamination, treatment delay, and systemic illness. 2 McLain et al and Duncan et al simply downscaled the original Gustilo–Anderson classification to apply to the hand. 3 4 More recently, Tulipan and Ilyas suggested that the Gustilo–Anderson system’s emphasis on wound size be abandoned in favor of a system that instead uses fracture location, wound contamination, soft tissue damage, and vascular viability to guide prognosis and management. 5 Clearly, there is a lack of consensus on open hand fracture classification; this uncertainty also extends to open hand fracture treatment.
A high percentage of hand fractures are open fractures, with studies reporting rates ranging from 34 to 68%. 6 Open hand fractures are particularly unique because they are frequently associated with a crush injury and extensively involve soft tissue. Despite these factors, anecdotal evidence has historically maintained that open hand fractures are less likely to become infected than open long bone fractures, perhaps due to the extensive vascularity of the hand. While reported infection rates have varied in recent studies of open hand fractures, they are still less than those seen in a meta-analysis of all open fractures. 5 7
Multiple researchers have attempted to delineate the appropriate timing of operative debridement after an open fracture. The unofficial 6-hour rule for the timing of surgical treatment is believed to be derived from Paul Leopold Friedrich’s research on guinea pigs, which found decreased infection and complication rates with debridement within 6 hours of the injury compared with debridement after 6 hours from the time of injury. 8 The 6-hour rule was further bolstered by Robson et al who demonstrated that 5.2 hours was the time necessary for wounds to develop greater than 10 5 bacterial colonies per gram of tissue, a bacterial concentration that is correlated with the development of infection. 9
Similarly, the unofficial 3-hour rule for antibiotics was derived from Patzakis and Wilkins’ analysis of more than 1,000 open fractures, which demonstrated increased infection rates in patients who received antibiotics more than 3 hours after the time of injury. 10 This study found that nearly one-half of infections occurred in open tibia fractures, but no other fracture location subgroups were described. 10
The purpose of our study was to investigate which factors impact infection rates after open hand fractures, including time to antibiotic administration and time to operative debridement. A secondary aim was to similarly evaluate factors that influence the development of nonunion after open hand fractures since infection can also contribute to nonunion. 1 Our hypothesis was that earlier antibiotic administration and earlier time to operative debridement would be associated with lower infection and nonunion rates.
Materials and Methods
We identified 1,090 patients with open hand fractures from January 1, 2012, through January 1, 2017, using a search of the International Classification of Disease 9th and 10th Revision codes for open hand fractures. Of these, 67 met inclusion criteria and had sufficient follow-up for one or both primary outcomes (infection or nonunion). Inclusion criteria included age of 18 years or older at the time of injury, presence of an open fracture of the phalanx, metacarpal, or carpal bones, and adequate information in the medical record regarding the time of injury, time of antibiotic administration, time of operative debridement, and clinical and radiological outcomes. Patients with insufficient documentation in the medical record and those who presented in a delayed manner or as a transfer from an outside hospital were excluded from the study, as were patients whose open fractures were traumatic amputations or required immediate revision amputation.
Institutional Review Board approval was sought prior to the beginning of the study. No external sources of funding were used for this study. All data was obtained from the Synthetic Derivative, our academic medical center’s database for clinical information extracted from the electronic medical record. Data in the Synthetic Derivative consists of deidentified clinical notes and reports from the medical record system and is unaltered from chart documentation other than the patient’s identifying characteristics.
Recorded patient variables included patient age and sex, chronic medical conditions (diabetes mellitus, cirrhosis, cancer), time and mechanism of injury, modified Gustilo–Anderson classification, time to antibiotic administration, time to operative debridement, and clinical and radiological follow-up time.
Outcome measures were the development of infection and fracture union. If the clinical documentation directly reported a Gustilo–Anderson classification, this classification was used. If the clinical documentation did not directly report a Gustilo–Anderson classification, review of available clinical documentation and radiological reports determined the classification. Mechanism of injury, wound contamination, and size of open wounds were obtained from the clinical notes. Fracture characteristics such as the presence of comminution were obtained from radiological reports. These factors were then used to classify each fracture, as described in the following. Definitions for Gustilo–Anderson classifications were taken from the work of McLain et al’s modification of the classic Gustilo–Anderson open long bone fracture classification system and are delineated in Table 1 . 3 11 Corresponding with McLain et al’s modified classification system, no distinction was made for A, B, or C subtypes of type III fractures. 3 Infection was determined by requirement of antibiotics and/or operative debridement to treat an infection or presumed infection as determined by the clinical team. Fracture union was determined by clinical notes, radiographic reports, and operative reports. To be included in this study, a minimum of 8 weeks of follow-up time was required.
Table 1. Modified Gustilo–Anderson open hand fracture classification system 3 11 .
| Type I | Tidy laceration less than 1 cm in length, without contamination, soft tissue crush, or comminution |
| Type II | Tidy laceration greater than 1 cm in length without contamination, soft tissue crush, stripping, or fracture comminution |
| Type III | Laceration greater than 10 cm in length or an untidy laceration, soft tissue crush, periosteal stripping, comminuted or segmental fracture, blast injuries, wound contamination, and all farming injuries |
Clinical Management
Patients were evaluated by the hand team, which is jointly covered by the orthopaedic surgery service and the plastic surgery service. Occasionally, the patient may have received antibiotics from the emergency department providers prior to evaluation by the hand team. The covering service determined the antibiotic regimen, but it was typically 2 g of intravenous cefazolin. Alternate antibiotic regimens included 900 mg of clindamycin, 825 mg of amoxicillin plus 125 mg of clavulanate, and 2 g of ampicillin plus 1 g of sulbactam. All antibiotics were redosed at appropriate intervals until operative debridement or discharge. The covering service also determined surgical care. Some patients were taken to the operating room on an urgent or semiurgent basis as determined by the attending surgeon and the patient’s specific clinical situation. A subset of patients received only an irrigation and closure procedure in the emergency department and did not require operative debridement or fixation. No protocol was used for surgical management; decisions regarding type of fixation and timing of operative intervention were determined by the attending surgeon. The attending surgeon also determined further management.
Statistical Analysis
The first primary outcome was the development of infection, defined as the requirement of antibiotics or surgical debridement. A secondary outcome was a documented diagnosis of nonunion based on clinical documentation, operative reports, or radiographic reports. Differences between groups (infection vs. no infection and nonunion versus union) in terms of patient demographics and clinical characteristics were assessed using Student’s t -test for continuous variables and chi-square test for categorical variables. An F -test was used to determine whether the data had equal or unequal variances. Continuous variables are reported as mean ± standard deviation. Bivariate linear regression analyses were used to determine the association between patient demographics and clinical characteristics and the presence of infection or a diagnosis of nonunion. Significance was considered to be p < 0.05.
Results
The final study population included 67 patients, with an age range of 19 to 80 years and an average age of 41.4 years. The study population included 82.1% males. Multiple open hand fractures were present in 24 (35.8%) of the patients for a total of 107 fractures. Patients were analyzed individually regardless of their number of fractures, and there was no significant difference in infection or union rates between patients with single versus multiple fractures. The study population had a minimum follow-up time of 8 weeks, with a median follow-up time of 14 weeks.
The time from injury to antibiotic administration ranged from 5 minutes to 20 hours, with a mean of 2.41 hours and a median of 1.62 hours. Antibiotics were administered within 3 hours of time of injury 87% of the time. Operative debridement was performed within 6 hours of time of injury for 25.4% of the patients, and 53.7% received operative debridement more than 6 hours after the time of injury. Approximately one-fifth (20.9%) of the patients did not require operative debridement and were treated with irrigation and closure in the emergency department. These patients were separated based on the timing of irrigation and closure into the aforementioned groups (within 6 hours of injury or more than 6 hours from injury).
Six of the 67 patients developed infection, with an overall infection rate of 9% of patients. Five of these six patients received antibiotics for superficial infections. One patient developed a necrotic wound that required amputation and subsequently developed a chronic infection with a suspected sinus tract to the necrotic bone that required a revision amputation. Six of the patients developed nonunion, with an overall nonunion rate of 9% of patients. Four of these six patients required an additional surgery to treat their nonunion. No patients developed both infection and nonunion.
Table 2 delineates the breakdown of infection and nonunion by the modified Gustilo–Anderson classification. None of the type I and II fractures became infected or developed nonunion, whereas 12.2% of the type III fractures developed infection and 12.2% developed nonunion. Type I and II fractures were combined together for statistical analysis because there was only one type I fracture. There was a statistically significant difference in infection rates and nonunion rates between type I/II and type III fractures ( p = 0.04).
Table 2. Infection and nonunion rates by the modified Gustilo – Anderson classification .
| Modified Gustilo–Anderson classification | Total | Infection, n (%) | Nonunion, n (%) | p -Value |
|---|---|---|---|---|
| Type I/II | 18 | 0 (0%) | 0 (0%) | |
| Type III | 49 | 6 (12.2%) | 6 (12.2%) | |
| 0.04 |
Table 3 compares patient demographics and clinical characteristics by diagnosis of infection. Patients who received antibiotics within 3 hours of injury demonstrated a lower infection rate, but this was not statistically significant. Notably, only nine patients received antibiotics beyond 3 hours, making it difficult to draw conclusions from this data subset. There was also no significant difference between infection rates based on time to operative debridement. Interestingly, patients who only underwent irrigation and closure in the emergency department had a lower infection rate than patients who required operative debridement, but this difference was not statistically significant. Patients with infection were not statistically different in terms of smoking status and chronic disease comorbidities, although smoking patients and patients with chronic disease did have higher rates of infection.
Table 3. Differences in patient demographics and clinical characteristics by infection.
| Characteristic | Total ( n = 67) | Infection ( n = 6) | No infection ( n = 61) | p -Value | Patients developing infection (%) |
|---|---|---|---|---|---|
| Time to antibiotic, hours | 0.99 | ||||
| <3 | 58 | 5 | 53 | 8.6 | |
| >3 | 9 | 1 | 8 | 11.1 | |
| Time to operative debridement, hours | 0.98 | ||||
| <6 | 28 | 3 | 25 | 10.7 | |
| >6 | 39 | 3 | 36 | 7.7 | |
| Smoker | 0.49 | ||||
| Yes | 25 | 4 | 21 | 16 | |
| No | 42 | 2 | 40 | 9.5 | |
| Chronic disease | 0.53 | ||||
| Yes | 9 | 2 | 7 | 22.2 | |
| No | 58 | 4 | 54 | 6.9 |
Table 4 compares patient demographics and clinical characteristics by diagnosis of nonunion. Patients who received antibiotics within 3 hours of injury demonstrated a lower nonunion rate, but this was not statistically significant. There was also no significant difference between nonunion rates based on time to operative debridement. Although both smoking patients and patients with chronic disease had higher rates of nonunion, these subgroups were not statistically different from nonsmoking patients or patients without chronic disease.
Table 4. Differences in patient demographics and clinical characteristics by nonunion.
| Characteristic | Total ( n = 67) | Nonunion ( n = 6) | Union ( n = 61) | p -Value | Patients developing nonunion (%) |
|---|---|---|---|---|---|
| Time to antibiotic, hours | 0.99 | ||||
| <3 | 58 | 5 | 53 | 8.6 | |
| >3 | 9 | 1 | 8 | 11.1 | |
| Time to operative debridement, hours | 0.98 | ||||
| <6 | 28 | 3 | 25 | 10.7 | |
| >6 | 39 | 3 | 36 | 7.7 | |
| Smoker | 0.93 | ||||
| Yes | 25 | 3 | 22 | 12 | |
| No | 42 | 3 | 39 | 7.1 | |
| Chronic disease | 0.53 | ||||
| Yes | 9 | 2 | 7 | 22.2 | |
| No | 58 | 4 | 54 | 6.9 |
The average time of antibiotic administration was not significantly different in patients developing infection or nonunion ( Table 5 ). The average time to operative debridement did not significantly differ for patients who developed infection or nonunion.
Table 5. Time to antibiotic administration and operative debridement from the time of injury.
| Time to antibiotic administration, hours | Time to operative procedure or irrigation/closure, hours | |
|---|---|---|
| Infection | 3.1 ± 4.0 | 85.5 ± 118.2 |
| No infection | 2.3 ± 2.7 | 52.5 ± 192.8 |
| p = 0.55 | p = 0.70 | |
| Nonunion | 1.9 ± 0.7 | 176.3 ± 285.7 |
| Union | 2.5 ± 3.0 | 43.5 ± 93.6 |
| p = 0.24 | p = 0.31 |
There was also no significant difference in rates of infection or nonunion in patients who received antibiotics in less than 3 hours and underwent debridement in less than 6 hours compared with those who received one or both of these interventions beyond these time periods ( Table 6 ).
Table 6. Comparison of patients who received antibiotics in less than 3 hours and were debrided in less than 6 hours (group A) and those who received 1 or both of these interventions beyond these times (group B).
| Infection | ||||
|---|---|---|---|---|
| Group | Total | Infection ( n = 6) | No infection ( n = 61) | p -Value |
| A | 27 | 3 | 24 | |
| B | 40 | 3 | 37 | |
| 0.97 | ||||
| Nonunion | ||||
| Group | Total |
Nonunion
( n = 6) |
Union
( n = 61) |
p -Value |
| A | 27 | 3 | 24 | |
| B | 40 | 3 | 37 | 0.97 |
Discussion
Neither time from injury to antibiotic administration nor time to operative debridement was significantly associated with an increased rate of infection or nonunion, refuting our original hypothesis. The factor associated with increased rates of infection was severity of injury as determined by the modified Gustilo–Anderson classification. Although not significant, patients requiring only irrigation and closure in the emergency department actually had slightly lower infection rates, perhaps reflecting a lower severity of injury. Type III fractures were also significantly more likely to develop nonunion than type I or II fractures.
We believe that this study adds to the expanding literature refuting the historical 6-hour rule for time to operative debridement, as no increased rate of infection was seen in patients who were debrided within this time frame. Moreover, patients who received antibiotics in less than 3 hours and underwent debridement within 6 hours did not have higher rates of fracture union than patients who were treated beyond these time frames.
Prior studies have investigated open fractures of the hand, distal radius, ulna, and humerus. While these studies report varying infection rates, most have concurred with this study’s finding that factors other than initial management are more influential on outcomes such as infection and nonunion. Swanson et al reported an infection rate of 6% in 116 patients with open hand fractures. 2 The presence of wound contamination and/or chronic disease comorbidities were the factors most associated with the development of infection. 2 As a result, these researchers suggested a novel open hand fracture classification scheme based on the presence of contamination and chronic illness. 2 Similarly, Duncan et al reported infections in 3.5% of open hand fractures (6 infections in 171 fractures). 4 Development of deep infection was associated with severity of injury and the presence of systemic illness. 4 Poor outcomes regarding range of motion were correlated with the severity of injury but not with systemic illness. 4 McLain et al reported an 11% infection rate in 143 patients with open hand fractures and likewise found increased infection rates with gross wound contamination and severity of injury as characterized by extensive soft tissue involvement or skeletal trauma. 3 In a retrospective review of 70 patients with open hand fractures, Ng et al found an infection rate of 11.4% and no significant impact of operative debridement timing on the development of infection. 12 These researchers did not compare specific timing of antibiotic administration but did find a significantly increased infection rate in patients who did not receive any antibiotics. 12
Previous work from our institution has found no association of timing of antibiotic administration and operative debridement with the development of infection or nonunion in open distal radius, ulna, and humerus fractures. 13 14 Similar to the results of this study, severity of injury was the factor most associated with the development of complications such as infection. 13 14
Strengths of our study include a sophisticated electronic medical record and deidentified clinical database, which provided detailed information on the timing of operative debridement and antibiotic administration. Our institution is a major trauma center and thus receives open fractures from a wide surrounding area. Because open hand fractures are jointly covered by both orthopaedic surgery and plastic surgery, this study likely represents a more balanced view on management options than those which are only managed by a single service.
Limitations of our study include its retrospective nature, which constrained the amount of information we could obtain. Due to our institution’s large referral area, many patients could not be included because of prior care received at an outside hospital or lack of sufficient follow-up at our institution. We also recognize that our patient cohort has a high rate of type III injuries, which likely reflects both the severity of injuries treated at our level 1 trauma center and our interpretation of McLain et al’s modification of the Gustilo–Anderson classification system. Many injuries that would have been classified as type II injuries based on wound size criteria alone were classified as type III open fractures, typically due to blast mechanism, comminution, or extensive soft tissue injury. It is important to note that our small sample size likely means that this study is underpowered and it is possible that a larger sample size could impact the statistical conclusions. However, due to the wide catchment of our trauma center, we believe that our results can likely be generalized to most patient populations.
Conclusion
Factors such as time to antibiotics and operative debridement were not associated with the development of infection or nonunion in this study. However, we do not mean to imply that these interventions should be delayed. Antibiotic administration is a simple, uncomplicated intervention that should be performed in the emergency department as soon as possible. Similarly, we continue to advocate for prompt, although not necessarily emergent, operative debridement or irrigation and closure for open hand fractures. Each patient’s clinical situation and comorbidities should be analyzed individually to determine the appropriate timing of debridement and type of fixation. However, we do note that our study and those previously mentioned do not support improved outcomes with operative debridement in less than 6 hours. Although we maintain that the optimal care of open hand fractures includes immediate antibiotic administration and opportune debridement, the data from this study and others suggest that injury severity rather than initial management will have the most substantial impact on outcomes.
Acknowledgments
The dataset used for the analyses described was obtained from Vanderbilt University Medical Center’s BioVU, which is supported by institutional funding, the 1S10RR025141–01 instrumentation award, and the CTSA grant UL1TR000445 from NCATS/NIH. The authors express deep gratitude to Ms. Julie Shelton for her hard work, contributions, and invaluable assistance in the design, coordination, and completion of this study.
Footnotes
Conflict of Interest None declared.
References
- 1.Mills L, Tsang J, Hopper G, Keenan G, Simpson A H. The multifactorial aetiology of fracture nonunion and the importance of searching for latent infectio. Bone Joint Res. 2016;5(10):512–519. doi: 10.1302/2046-3758.510.BJR-2016-0138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Swanson T V, Szabo R M, Anderson D D. Open hand fractures: prognosis and classificatio. J Hand Surg Am. 1991;16(01):101–107. doi: 10.1016/s0363-5023(10)80021-8. [DOI] [PubMed] [Google Scholar]
- 3.McLain R F, Steyers C, Stoddard M. Infections in open fractures of the han. J Hand Surg Am. 1991;16(01):108–112. doi: 10.1016/s0363-5023(10)80022-x. [DOI] [PubMed] [Google Scholar]
- 4.Duncan R W, Freeland A E, Jabaley M E, Meydrech E F. Open hand fractures: an analysis of the recovery of active motion and of complication. J Hand Surg Am. 1993;18(03):387–394. doi: 10.1016/0363-5023(93)90080-M. [DOI] [PubMed] [Google Scholar]
- 5.Tulipan J E, Ilyas A M. Open fractures of the hand: review of pathogenesis and introduction of a new classification syste. Hand Clin. 2018;34(01):1–7. doi: 10.1016/j.hcl.2017.09.013. [DOI] [PubMed] [Google Scholar]
- 6.Chow S P, Pun W K, So Y C et al. A prospective study of 245 open digital fractures of the han. J Hand Surg [Br] 1991;16(02):137–140. doi: 10.1016/0266-7681(91)90162-h. [DOI] [PubMed] [Google Scholar]
- 7.Schenker M L, Yannascoli S, Baldwin K D, Ahn J, Mehta S. Does timing to operative debridement affect infectious complications in open long-bone fractures? A systematic revie. J Bone Joint Surg Am. 2012;94(12):1057–1064. doi: 10.2106/JBJS.K.00582. [DOI] [PubMed] [Google Scholar]
- 8.Friedrich P L. Die aseptische Versorgung frischer Wunden, unter Mittheilung von Thier-Versuchen über die Auskeimungszeit von Infection serregern in frischen Wunde. Arch für Klinsche Chirugie. 1898;57:288–310. [Google Scholar]
- 9.Robson M C, Duke W F, Krizek T J. Rapid bacterial screening in the treatment of civilian wound. J Surg Res. 1973;14(05):426–430. doi: 10.1016/0022-4804(73)90049-8. [DOI] [PubMed] [Google Scholar]
- 10.Patzakis M J, Wilkins J. Factors influencing infection rate in open fracture wound. Clin Orthop Relat Res. 1989;14(243):36–40. [PubMed] [Google Scholar]
- 11.Gustilo R B, Anderson J T. Prevention of infection in the treatment of one thousand and twenty-five open fractures of long bones: retrospective and prospective analyse. J Bone Joint Surg Am. 1976;58(04):453–458. [PubMed] [Google Scholar]
- 12.Ng T, Unadkat J, Bilonick R A, Wollstein R. The importance of early operative treatment in open fractures of the finger. Ann Plast Surg. 2014;72(04):408–410. doi: 10.1097/SAP.0b013e318264fcd5. [DOI] [PubMed] [Google Scholar]
- 13.Cash B M, Zumsteg J W, Lee D H et al. Factors influencing infection and nonunion rates after open fractures of the humeru. J Hand Surg Am. 2018;43(09):S25–S26. doi: 10.1016/j.jhsa.2014.02.008. [DOI] [PubMed] [Google Scholar]
- 14.Zumsteg J W, Molina C S, Lee D H, Pappas N D. Factors influencing infection rates after open fractures of the radius and/or uln. J Hand Surg Am. 2014;39(05):956–961. doi: 10.1016/j.jhsa.2014.02.008. [DOI] [PubMed] [Google Scholar]