Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Scharfenberger et al J Orthop Trauma ! Volume 31, Number 3, March 2017
and would not lead to an increase in the development of either requiring bone grafting after debridement, grossly con-
deep infection or nonunion relative to delayed wound closure. taminated wounds, fractures resulting from penetrating
trauma (eg, gunshot wounds, and stabbings), pathologic
fractures, or patients unfit to undergo anesthetic. The most
PATIENTS AND METHODS common reasons for exclusion were higher grade fractures
Design and Setting [n = 29 (29%)] and requiring delayed closure [n = 28
(28%)] (Fig. 1).
Between 2009 and 2013, 183 subjects were screened
for enrollment at one Level 1 trauma center; 83 subjects with
84 open fractures were enrolled and followed for at least 1 Intervention
year postoperatively (Fig. 1). All participating surgeons were Initial surgical debridement and fracture fixation with
trauma fellowship-trained surgeons with a minimum of 5 copious irrigation (3 liters or more) and debridement of soft
years of experience. Signed informed consent was provided tissues and contaminated bone were undertaken. Surgical
at index hospitalization by participants. The regional health fixation was at the surgeons’ discretion. At the time of sur-
ethics board approved this study (Pro00009272). gery, eligible subjects underwent primary wound closure after
surgical fixation. Tetanus prophylaxis was provided when
Selection Criteria immunization status was either unclear or not up to date
Potential subjects met the following criteria: skeletal and a standardized antibiotic regimen that was approved by
maturity (as seen on radiographs), long bone open fractures Orthopaedic Surgery, Pharmacy, and Infectious Diseases was
(humerus, radius/ulna, femur, and tibia/fibula), presenting for followed for all patients (see Appendix, Supplemental
initial surgical debridement, and patient or proxy respondent Digital Content 1, http://links.lww.com/BOT/A830). All
able to provide consent. Open fractures were defined as patients underwent a 48-hour postdebridement wound check
fractures where the bone was exposed to the environment by the surgeon or their designate.
through a wound to skin and soft tissue. Those with Gustilo For patients who initially received treatment at non-
grade 1, 2, or 3A whose wounds/fractures were deemed clean surgical centers before transfer to the surgical site, medical
after initial debridement were considered eligible. Although the care included medical stabilization, initiation of antibiotics,
Gustilo classification was initially developed for tibial fractures and nonoperative stabilization of the fracture(s).
and has limitations in its reliability,13 it has since been extended
to include other open fractures of the long bones, consistent Data Collection
with commonly accepted practice and the initial grade 3 clas- Patient characteristics (eg, age, sex, and comorbidities),
sification has been further categorized as 3A, 3B, or 3C.14 injury information (eg, Gustilo grade, fracture site; timing of
The following criteria were reason for exclusion: antibiotic administration, and transfusion requirement), and
patients with Gustilo grade 3B or 3C fractures, patients health services information (eg, time from injury to initial
122 | www.jorthotrauma.com Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma ! Volume 31, Number 3, March 2017 Time to Initial Operative Treatment
surgical management, and length of hospital stay) were subjects in the primary closure cohort, matching was not
recorded. Research associates recorded preexisting medical attempted on any further variables. Grade of fracture was
conditions including smoking status at the time of fracture. chosen over fixation method for matching as grade of fracture
Surgeons completed wound classification forms15 after the could affect fixation method selected.
first surgery so that grade of fracture was ascertained before
any outcome assessment. Fractures were classified as Analysis
humeral, radio/ulnar, femoral, or tibia/fibular, in location. Descriptive analyses (mean, SDs, medians, interquartile
No open pelvic fractures were included. Upper extremity ranges, frequencies, and proportions) were undertaken with the
fractures (humeral and radio/ulnar) were collapsed into a sin- primary cohort for all recorded variables including the incidence
gle category for analysis. of deep infection and nonunion. The number and proportion of
Attending surgeons evaluated subjects using standard- subjects who developed deep infection and/or nonunion were
ized data forms until the fracture(s) healed. Research also examined by Gustilo grade and fracture location.
associates performed phone interviews and chart reviews at For the matched analysis, where there were multiple
least 1 year postfracture to confirm infection and nonunion matches, the first match from the list was selected for
outcomes as well as any reoperations. Subjects needed to analysis. At the time of matching, evaluators were blinded
complete the 1-year telephone interview or have clinical to outcomes in both cohorts. McNemar tests were used for the
follow-up of at least 90 days after surgery with a definitive matched analyses.
clinical outcome (ie, healed fracture) recorded to provide All analyses were performed using the Statistical
adequate follow-up data. Package for the Social Sciences (SPSS) version 22.0 (SPSS
Inc, Chicago, IL) using 2-tailed tests and a significance level
Main Outcome Measurement of a = 0.05.
Deep infection was defined as infection requiring
unplanned surgical debridement and/or sustained antibiotic
therapy after definitive wound closure. Cellulitis and pin tract RESULTS
infections alone were not considered indicative of deep Demographics
wound infection, but these were treated with the appropriate
More subjects in the primary closure cohort were male
antibiotics at the surgeons’ discretion and outcomes recorded.
[n = 52 (62%)] and the median age was 45.8 (minimum 17,
Nonunions were defined as unplanned surgical intervention
maximum 88) years (Table 1). Almost half of subjects [n = 38
after definitive wound closure or incomplete radiographic
(45%)] sustained other injuries, but only 1 subject sustained
healing 1-year postfracture.
multiple open fractures. Most injuries occurred in motor vehicle
Outcome Ascertainment accidents [n = 38 (45%)] or falls [n = 34 (41%)]. Assaults [n = 6
(7%)] and crush injuries [n = 6 (7%)] accounted for the remain-
Outcomes were classified using data from standardized
ing mechanisms of injury. Most subjects had fewer than 2
outpatient clinic, telephone interview, and chart review forms.
comorbidities [n = 66 (74%)]. Eight (10%) subjects had diabe-
Outpatient clinic data forms completed by the attending
tes mellitus, and 45 (54%) were current smokers (Table 1).
surgeon with infection and nonunion status identified were
the primary source of information. One-year interview Fracture Characteristics
information and chart reviews/radiographic assessment were
Upper extremity fractures occurred in 38 (45%) sub-
used to confirm outcomes and to assess for any late-occurring
jects and tibia/fibular fractures occurred in 36 (43%) subjects
complications (ie, infections). Where discrepancies were
with the remaining 10 (12%) fractures occurring in the femur.
found, the most recent source of either reoperation/radiographic
Gustilo grade 1 and 2 fractures occurred equally [n = 35
review or surgeon record of outcome was considered
the definitive outcome. All infections and nonunions were
confirmed through clinical records.
TABLE 1. Characteristics of 83 Subjects With Open Fractures
Delayed Closure Cohort of a Long Bone
A matched-series analysis was performed using cohort Subject Characteristics
data of subjects who were enrolled in a study of 736 subjects Total no. subjects 83
with 791 fractures that examined the impact of time to Median age (IQR) 45.8 (27.0–65.3)
antibiotic administration and surgery on development of deep Sex—male, % 52 (62)
infection and nonunion.11,12 This previous study was carried Comorbidities, %
out between 2001 and 2009 and used the same methods, None 40 (48)
evaluation, and definition of outcomes. 1–2 conditions 22 (26)
To be eligible for the matched analyses, subjects from 3 or more conditions 22 (26)
the initial cohort had to have undergone more than 1 surgical Diabetes mellitus* 8 (10)
debridement at index hospitalization and to have undergone Current Smoker at the time of injury, 45 (54)
surgery at the same trauma center (n = 370 patients). To %
match cases, subjects were grouped by age, sex, fracture *Counted in the Comorbidities.
location, and Gustilo grade. Because of the small number of
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved. www.jorthotrauma.com | 123
Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Scharfenberger et al J Orthop Trauma ! Volume 31, Number 3, March 2017
124 | www.jorthotrauma.com Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma ! Volume 31, Number 3, March 2017 Time to Initial Operative Treatment
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved. www.jorthotrauma.com | 125
Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Scharfenberger et al J Orthop Trauma ! Volume 31, Number 3, March 2017
with upper extremity.11 Upper extremity fractures accounted a definitive randomized trial that focuses on primary closure
for almost half of our enrolled subjects. With the low number in lower extremity fractures is likely required to answer this
of adverse outcomes in our cohort, we were unable to do any question, this study supports the use of primary wound
in-depth analysis of how fracture location may affect out- closure in open fracture as it appears to reduce infection
comes when primary closure is used. Future work on primary and nonunion rates compared with delayed wound closure.
closure in open fractures should focus on the lower extremity
to determine whether there are other clinical factors that affect
outcomes when primary closure is used. REFERENCES
In our matched-series analysis with similar subjects 1. Schenker ML, Yannascoli S, Baldwin KD, et al. Does timing to operative
who received delayed wound closure, we were also unable to debridement affect infectious complications in open long-bone fractures?
A systematic review. J Bone Joint Surg Am. 2012;94:1057–1064.
match on wound cleanliness between cohorts. We also did not 2. Hampton OP, Jr. Basic principles in management of open fractures.
control for segment of bone fractured and note that there are JAMA. 1955;159:417–419.
issues with the reliability of assigning Gustilo grades.13 Thus, 3. Trueta J. The treatment of war fractures by the closed method. Proc R
although primary closure was not commonly used in the ear- Soc Med. 1939;33:65–74.
lier cohort, our results should be interpreted cautiously as 4. Trueta J. Reflections on the past and present treatment of war wounds
and fractures. Mil Med. 1976;141:255–258.
clinical biases may exist between cohorts for which matching 5. Benson DR, Riggins RS, Lawrence RM, et al. Treatment of open frac-
did not control. Because our study was relatively small with tures: a prospective study. J Trauma. 1983;23:25–30.
fewer than 100 subjects, we did not use propensity scoring, 6. DeLong WJ, Born CT, Wei SY, et al. Aggressive treatment of 119 open
but instead performed direct matching on a limited number of fracture wounds. J Trauma. 1999;46:1049–1054.
7. Hohmann E, Tetsworth K, Radziejowski MJ, et al. Comparison of
variables to maximize the number of pairs achieved between delayed and primary wound closure in the treatment of open tibial frac-
the 2 cohorts. However our results are consistent with the tures. Arch Orthop Traum Surg. 2007;127:131–136.
current evidence on this topic.5–9 8. Jenkinson RJ, Kiss A, Johnson S, et al. Delayed wound closure in-
The surgeons involved in this study are experienced creases deep-infection rate associated with lower-grade open fractures:
trauma surgeons working at a high volume level 1 trauma a propensity-matched cohort study. J Bone Joint Surg Am. 2014;96:
380–386.
center with a large geographic catchment area. This may limit 9. Moola FO, Carli A, Berry GK, et al. Attempting primary closure for all
the applicability of our findings to smaller centers with lower open fractures: the effectiveness of an institutional protocol. Can J Surg.
patient volumes and less experienced surgeons who serve 2014;57:E82–E88.
a primarily urban-based populace, particularly because the 10. Carsenti-Etesse H, Doyon F, Desplaces N, et al. Epidemiology of bac-
terial infection during management of open leg fractures. Eur J Clin
choice of primary or delayed wound closure was to some Microbiol Infect Dis. 1999;18:1999.
degree at the surgeon’s discretion. However, our results are 11. Weber D, Dulai SK, Bergman J, et al. Time to initial operative treatment
similar to those recently reported by other high volume following open fracture does not impact development of deep infection:
trauma centers.8,9 a prospective cohort study of 736 subjects. J Orthop Trauma. 2014;28:
613–619.
Finally, we did not consider any economic impact of 12. Westgeest J, Weber D, Dulai SK, et al. Factors associated with develop-
primary versus delayed wound closure. We found that ment of nonunion or delayed healing after an open long bone fracture:
subjects undergoing primary closure stayed a similar length a prospective cohort study of 736 subjects. J Orthop Trauma. 2016;30:
of time in hospital as those who underwent delayed wound 149–155.
closure. We did not consider the cost of additional operating 13. Brumback RJ, Jones AL. Interobserver agreement in the classification
of open fractures of the tibia. The results of a survey of two hundred
room visits associated with a delayed wound closure and forty-five orthopaedic surgeons. J Bone Joint Surg Am. 1994;76:
approach, which is an important economic consideration, 1162–1166.
particularly when primary wound closure may lead to at least 14. Gustilo RB, Mendoza RM, Williams DN. Problems in the management
equivocal and potentially superior patient outcomes relative of type III (severe) open fractures: a new classification of type III open
to delayed wound closure. Economic impact should also be fractures. J Trauma. 1984;24:742–746.
15. Gustilo RB, Anderson JT. Prevention of infection in the treatment of one
considered as an important outcome in any future trial. thousand and twenty-five open fractures of long bones: retrospective and
prospective analyses. J Bone Joint Surg Am. 1976;58:453–458.
16. Marsh JL, Slongo TF, Agel J, et al. Fracture and Dislocation Classifica-
CONCLUSIONS tion Compendium - 2007: Orthopaedic Trauma Association Classifica-
In summary, the low rate of deep infection and tion, Database and Outcomes Committee. J Orthop Trauma. 2007;21
Supplement 10 pp: S1–S163.
nonunion with primary wound closure for clean Gustilo 17. Lenarz CJ, Watson JT, Moed BR, et al. Timing of wound closure in open
grade 3A or less fractures, lends credence to the practice of fractures based on cultures obtained after debridement. J Bone Joint Surg
primary closure in appropriately selected subjects. Although Am. 2010;92:1921–1926.
126 | www.jorthotrauma.com Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.