ORIGINAL ARTICLE
Primary Wound Closure After Open Fracture: A Prospective
Cohort Study Examining Nonunion and Deep Infection
Angela V. Scharfenberger, MD, FRCS,* Khaled Alabassi, MD,* Stephanie Smith, MD, FRCP,†
Donald Weber, MD, FRCS,* Sukhdeep K. Dulai, MD, MSc, FRCS,* Joseph W. Bergman, MD, FRCS,*
and Lauren A. Beaupre, PT, PhD*‡
Objectives: Determine the proportion of subjects developing deep
infection or nonunion after primary wound closure of open fractures
(humerus, radius/ulna, femur, and tibia/fibula). Secondarily,
a matched-series analysis compared outcomes with subjects who
underwent delayed wound closure.
Design: Prospective cohort between 2009 and 2013 of subjects
undergoing primary closure.
Setting: Trauma center.
Participants: Eighty-three (84 fractures) subjects were enrolled.
Eighty-two (99%) subjects (83 fractures) provided follow-up data.
Matching (age, sec, fracture location, and grade) was performed
using study data of delayed wound closure undertaken at the same
center between 2001 and 2009 (n = 68 matched subjects).
Intervention: Primary wound closure occurred when the fracture
grade was Gustilo grade 3A or lower and the wound deemed clean at
initial surgery. Standardized evaluations occurred until the fracture
(s) healed; phone interviews and chart reviews were also undertaken
at 1 year.
Main Outcome Measurements: Deep infection is defined as
infection requiring unplanned surgical debridement and/or sustained
antibiotic therapy after wound closure; nonunion is defined as
unplanned surgical intervention after definitive wound closure or
incomplete radiographic healing 1-year after fracture.
Accepted for publication November 4, 2016.
From the Departments of *Surgery; †Infectious Diseases; and ‡Physical
Therapy, University of Alberta, Edmonton, AB, Canada.
This study was funded through unrestricted research grants from the
Edmonton Civic Employees Charitable Foundation and the Edmonton
Orthopaedic Research Committee.
L. A. Beaupre receives salary support from the Canadian Institutes for Health
Research as a New Investigator (Patient Oriented Research) and Alberta
Innovates—Health Solutions as a Population Health Investigator. The
remaining authors report no conflict of interest.
Presented as a poster at the Annual Meeting of the Orthopaedic Trauma
Association, October 7–10, 2015, San Diego, CA.
Supplemental digital content is available for this article. Direct URL citations
appear in the printed text and are provided in the HTML and PDF versions
of this article on the journal’s Web site (www.jorthotrauma.com).
This study was approved by the regional health ethics board at the University
of Alberta (Pro00009272).
Reprints: Lauren A. Beaupre, PT, PhD, 2-50 Corbett Hall, University of Alberta,
Edmonton, AB T6G 2G4, Canada (e-mail: lauren.beaupre@ualberta.ca).
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/BOT.0000000000000751
J Orthop Trauma ! Volume 31, Number 3, March 2017
Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Results: Three (4%) subjects had deep infections, whereas 10
(12%) subjects developed nonunion in the primary closure cohort. In
the matched analyses [n = 68 pairs; (136 subjects)], the primary
closure cohort had fewer deep infections [n = 3 (4%) vs. n = 6
(9%)] and nonunions [n = 9 (13%) vs. n = 19 (29%)] than the
delayed closure cohort (P , 0.001).
Conclusions: Primary wound closure after an open fracture
appears acceptable in appropriately selected patients and may reduce
the risk of deep infection and nonunion compared with delayed
closure; a definitive randomized trial is needed.
Key Words: open fracture, wound management, deep infection,
nonunion
Level of Evidence: Prognostic Level II. See Instructions for
Authors for a complete description of levels of evidence.
(J Orthop Trauma 2017;31:121–126)
INTRODUCTION
Open fractures of long bones require timely medical and
surgical management. Delayed wound closure was the accepted
approach for several decades to allow multiple debridements as
a measure to prevent deep infection.1 This approach arose from
treatment approaches used in military conflicts throughout the
20th century, including the Korean and Vietnam wars, where
there was a concern with deep infection, particularly infection
caused by clostridia species or other anaerobic organisms.2–4
More recently, a limited number of small studies and 1
larger cohort study with propensity score matching have
reported low infection rates when using a primary wound
closure approach.5–9 Primary wound closure offers some poten-
tial advantages in that immediate wound closure may protect
against hospital-acquired infections (nosocomial infections)10
and also may reduce the number of surgeries required, which
is advantageous both for patients and the health care system.
The primary study objective was to determine the
proportion of subjects who developed either a deep infection
or nonunion after primary wound closure of an open long bone
fracture. Secondarily, we compared outcomes using a matched-
series analysis of these outcomes (deep infection and nonunion)
with an historic cohort of subjects with long bone fractures who
had undergone delayed wound closure in a previous study at the
same trauma center.11,12 We hypothesized that primary wound
closure in appropriately selected subjects would be acceptable
www.jorthotrauma.com | 121
Scharfenberger et al
and would not lead to an increase in the development of either
deep infection or nonunion relative to delayed wound closure.
PATIENTS AND METHODS
Design and Setting
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
year postoperatively (Fig. 1). All participating surgeons were
trauma fellowship-trained surgeons with a minimum of 5
years of experience. Signed informed consent was provided
at index hospitalization by participants. The regional health
ethics board approved this study (Pro00009272).
Selection Criteria
Potential subjects met the following criteria: skeletal
maturity (as seen on radiographs), long bone open fractures
(humerus, radius/ulna, femur, and tibia/fibula), presenting for
initial surgical debridement, and patient or proxy respondent
able to provide consent. Open fractures were defined as
fractures where the bone was exposed to the environment
through a wound to skin and soft tissue. Those with Gustilo
grade 1, 2, or 3A whose wounds/fractures were deemed clean
after initial debridement were considered eligible. Although the
Gustilo classification was initially developed for tibial fractures
and has limitations in its reliability,13 it has since been extended
to include other open fractures of the long bones, consistent
with commonly accepted practice and the initial grade 3 clas-
sification has been further categorized as 3A, 3B, or 3C.14
The following criteria were reason for exclusion:
patients with Gustilo grade 3B or 3C fractures, patients
FIGURE 1. Flowchart of patient screening,
enrollment, and follow-up. Editor’s Note:
A color image accompanies the online
version of this article.
122 | www.jorthotrauma.com
Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma ! Volume 31, Number 3, March 2017
requiring bone grafting after debridement, grossly con-
taminated wounds, fractures resulting from penetrating
trauma (eg, gunshot wounds, and stabbings), pathologic
fractures, or patients unfit to undergo anesthetic. The most
common reasons for exclusion were higher grade fractures
[n = 29 (29%)] and requiring delayed closure [n = 28
(28%)] (Fig. 1).
Intervention
Initial surgical debridement and fracture fixation with
copious irrigation (3 liters or more) and debridement of soft
tissues and contaminated bone were undertaken. Surgical
fixation was at the surgeons’ discretion. At the time of sur-
gery, eligible subjects underwent primary wound closure after
surgical fixation. Tetanus prophylaxis was provided when
immunization status was either unclear or not up to date
and a standardized antibiotic regimen that was approved by
Orthopaedic Surgery, Pharmacy, and Infectious Diseases was
followed for all patients (see Appendix, Supplemental
Digital Content 1, http://links.lww.com/BOT/A830). All
patients underwent a 48-hour postdebridement wound check
by the surgeon or their designate.
For patients who initially received treatment at non-
surgical centers before transfer to the surgical site, medical
care included medical stabilization, initiation of antibiotics,
and nonoperative stabilization of the fracture(s).
Data Collection
Patient characteristics (eg, age, sex, and comorbidities),
injury information (eg, Gustilo grade, fracture site; timing of
antibiotic administration, and transfusion requirement), and
health services information (eg, time from injury to initial
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
J Orthop Trauma ! Volume 31, Number 3, March 2017
surgical management, and length of hospital stay) were
recorded. Research associates recorded preexisting medical
conditions including smoking status at the time of fracture.
Surgeons completed wound classification forms15 after the
first surgery so that grade of fracture was ascertained before
any outcome assessment. Fractures were classified as
humeral, radio/ulnar, femoral, or tibia/fibular, in location.
No open pelvic fractures were included. Upper extremity
fractures (humeral and radio/ulnar) were collapsed into a sin-
gle category for analysis.
Attending surgeons evaluated subjects using standard-
ized data forms until the fracture(s) healed. Research
associates performed phone interviews and chart reviews at
least 1 year postfracture to confirm infection and nonunion
outcomes as well as any reoperations. Subjects needed to
complete the 1-year telephone interview or have clinical
follow-up of at least 90 days after surgery with a definitive
clinical outcome (ie, healed fracture) recorded to provide
adequate follow-up data.
Main Outcome Measurement
Deep infection was defined as infection requiring
unplanned surgical debridement and/or sustained antibiotic
therapy after definitive wound closure. Cellulitis and pin tract
infections alone were not considered indicative of deep
wound infection, but these were treated with the appropriate
antibiotics at the surgeons’ discretion and outcomes recorded.
Nonunions were defined as unplanned surgical intervention
after definitive wound closure or incomplete radiographic
healing 1-year postfracture.
Outcome Ascertainment
Outcomes were classified using data from standardized
outpatient clinic, telephone interview, and chart review forms.
Outpatient clinic data forms completed by the attending
surgeon with infection and nonunion status identified were
the primary source of information. One-year interview
information and chart reviews/radiographic assessment were
used to confirm outcomes and to assess for any late-occurring
complications (ie, infections). Where discrepancies were
found, the most recent source of either reoperation/radiographic
review or surgeon record of outcome was considered
the definitive outcome. All infections and nonunions were
confirmed through clinical records.
Delayed Closure Cohort
A matched-series analysis was performed using cohort
data of subjects who were enrolled in a study of 736 subjects
with 791 fractures that examined the impact of time to
antibiotic administration and surgery on development of deep
infection and nonunion.11,12 This previous study was carried
out between 2001 and 2009 and used the same methods,
evaluation, and definition of outcomes.
To be eligible for the matched analyses, subjects from
the initial cohort had to have undergone more than 1 surgical
debridement at index hospitalization and to have undergone
surgery at the same trauma center (n = 370 patients). To
match cases, subjects were grouped by age, sex, fracture
location, and Gustilo grade. Because of the small number of
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Time to Initial Operative Treatment
subjects in the primary closure cohort, matching was not
attempted on any further variables. Grade of fracture was
chosen over fixation method for matching as grade of fracture
could affect fixation method selected.
Analysis
Descriptive analyses (mean, SDs, medians, interquartile
ranges, frequencies, and proportions) were undertaken with the
primary cohort for all recorded variables including the incidence
of deep infection and nonunion. The number and proportion of
subjects who developed deep infection and/or nonunion were
also examined by Gustilo grade and fracture location.
For the matched analysis, where there were multiple
matches, the first match from the list was selected for
analysis. At the time of matching, evaluators were blinded
to outcomes in both cohorts. McNemar tests were used for the
matched analyses.
All analyses were performed using the Statistical
Package for the Social Sciences (SPSS) version 22.0 (SPSS
Inc, Chicago, IL) using 2-tailed tests and a significance level
of a = 0.05.
RESULTS
Demographics
More subjects in the primary closure cohort were male
[n = 52 (62%)] and the median age was 45.8 (minimum 17,
maximum 88) years (Table 1). Almost half of subjects [n = 38
(45%)] sustained other injuries, but only 1 subject sustained
multiple open fractures. Most injuries occurred in motor vehicle
accidents [n = 38 (45%)] or falls [n = 34 (41%)]. Assaults [n = 6
(7%)] and crush injuries [n = 6 (7%)] accounted for the remain-
ing mechanisms of injury. Most subjects had fewer than 2
comorbidities [n = 66 (74%)]. Eight (10%) subjects had diabe-
tes mellitus, and 45 (54%) were current smokers (Table 1).
Fracture Characteristics
Upper extremity fractures occurred in 38 (45%) sub-
jects and tibia/fibular fractures occurred in 36 (43%) subjects
with the remaining 10 (12%) fractures occurring in the femur.
Gustilo grade 1 and 2 fractures occurred equally [n = 35
TABLE 1. Characteristics of 83 Subjects With Open Fractures
of a Long Bone
Subject Characteristics
Total no. subjects 83
Median age (IQR) 45.8 (27.0–65.3)
Sex—male, % 52 (62)
Comorbidities, %
None 40 (48)
1–2 conditions 22 (26)
3 or more conditions 22 (26)
Diabetes mellitus* 8 (10)
Current Smoker at the time of injury, 45 (54)
%
*Counted in the Comorbidities.
www.jorthotrauma.com | 123
Scharfenberger et al J Orthop Trauma ! Volume 31, Number 3, March 2017

(42%) each] with only 14 (16%) Gustilo grade 3A fractures (2

TABLE 2. Injury Characteristics of 82 Subjects With 83 Long
upper extremity, 8 tibia/fibula, and 3 femur) meeting inclu-
Bone Compound Fractures and Known Outcomes*
sion criteria.
The OTA/AO fracture classification was also completed Gustilo Not
Grade Infected Infected Healed Nonunion
for 81 (96%) fractures.16 Of these, seven (8%) upper extrem-
ity and 4 (5%) tibia/fibular fractures were proximal, whereas Fracture location
19 (22%) upper extremity, 9 (11%) tibia/fibular, and 5 (6%) Upper extremity 1 20 1† 20 1†
femoral fractures were distal. Diaphyseal fractures accounted 2 14 0 11 3
for 10 (12%) upper extremity, 9 (11%) tibia/fibular, and 3A 2 0 1 1
4 (5%) femoral fractures. Malleolar fractures occurred in Tibia/fibula 1 11 0 11 0
10 (12%) tibia/fibular fractures. Four (5%) fractures were 2 15 1 13 3*
segmental; 1 occurred in the upper extremity and 3 were 3A 8 1 9 0
tibia/fibula fractures. Femur 1 3 0 3 0
For fracture fixation, most subjects [n = 59 (70%)] 2 4 0 3 1
subjects received open reduction and internal fixation (ie, 3A 3 0 2 1
plates and screws or screws only), 23 (27%) subjects received *1 Subject had concurrent deep infection.
an intramedullary nail (reamed or unreamed), and 2 (2%) †Same subject.
received external fixation. For the 2 subjects managed with
external fixation (1 upper extremity and 1 tibial/fibula frac-
ture), external fixation was used as the definitive fracture There was no difference in union or infection outcomes
fixation and the wound was closed at the time of the index between matched and unmatched subjects (P = 1.0).
surgery. Twenty-four (29%) subjects received transfusions. In the matched analyses, mechanism of injury and
The median time to initial surgery was 10.75 hours (h) [in- associated injuries were similar (P . 0.08). The delayed
terquartile range (IQR) 7.94–15.44 hours], whereas the group reported more comorbidities, but the primary closure
median time to antibiotic administration was 3.25 hours cohort had more subjects with diabetes mellitus (Table 3).
(IQR 2.0–7.0 hours). The median hospital length of stay The median time to operative management and antibiotic
was 5.5 days (IQR 3.0–11.0) for the 83 primary closure administration was similar between cohorts as was the
participants. median length of stay in hospital (Table 3). There was also
no difference in fixation methods for the femoral or tibial
Follow-up fractures (P . 0.50) between the 2 cohorts, but there was
Overall, 70 (84%) subjects (71 fractures) completed the a shift to use more intramedullary nails for humeral fractures
1-year interview, and 82 (99%) subjects (83 fractures) and less external fixation of radial/ulnar fractures in the
completed clinic follow-up of greater than 90 days that primary closure cohort relative to the delayed closure cohort
allowed outcome ascertainment of healing or infection. (P = 0.001).
Radiographs were followed to healing or 1 year, whichever Overall, the primary closure cohort had fewer deep
came first, in all subjects except for 1 subject who was infections [n = 3 (4%) vs. n = 6 (9%)] than the delayed
deceased within 1 year of his fracture because of causes closure (P , 0.001; McNemar test for matched data). In
unrelated to his open fracture. The healing status of his the primary closure, 2 infections were in the tibia/fibula and
fracture was unknown. 1 was in the upper extremity, whereas in the delayed closure
cohort, 5 of the infections occurred in the tibia/fibula and 1
Primary Closure Cohort Outcomes occurred in the upper extremity.
Three (4%) primary closure subjects had deep infec- For nonunions, a similar pattern emerged with fewer
tions of which 2 were tibial/fibular fractures and 1 was an nonunions in the primary closure cohort [n = 9 (13%) vs. n =
upper extremity fracture. Ten (12%) subjects developed 19 (29%)] than the delayed closure cohort (P , 0.001;
nonunion—5 upper extremity, 3 tibia/fibular, and 2 femoral McNemar test for matched data). Two of the nonunions in
fractures. Two of the 3 subjects with deep infection also the primary closure cohort also had deep infections (1 tibia/
developed nonunion, 1 upper extremity and 1 tibial/fibular fibula and 1 upper extremity); in addition, there were 4 upper
fracture (Table 2). One subject failed primary closure and extremity, 2 tibia/fibula, and 1 femur nonunion. In the de-
required a second debridement at 48 hours because of a sus- layed closure cohort, all 6 of the fractures with deep infec-
pected superficial wound infection while in hospital; this frac- tions also had nonunion. In addition, there were 6 upper
ture subsequently healed without deep infection. extremity, 6 tibia/fibula, and 1 femur nonunion in the delayed
closure cohort.
Matched Analyses
Sixty-eight pairs (n = 136 subjects; 81% match rate)
were created that matched on all 4 variables (age, sex, fracture DISCUSSION
location, and Gustilo grade). Unmatched subjects from the In our study of 83 subjects with 84 open fractures with
primary closure cohort [n = 16 (19%)] were more likely to a Gustilo grade 3A or less who underwent primary wound
be older (P , 0.001) females (P = 0.009) with lower grade closure, we found low rates of both infection (4%) and
Gustilo (P = 0.009) upper extremity fractures (P = 0.009). nonunion (12%). Our rate of infection is similar to or lower

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

delayed closure may increase the risk for hospital-acquired in-

TABLE 3. Comparison of Primary Closure and Delayed
fections10 and delay mobilization and discharge from hospital.
Closure Participants Matched on Age, Sex, Grade, and
Location of Fracture One small randomized study5 and a limited number case
series or cohort studies have consistently reported similar re-
Primary Delayed
Closure Closure P
sults to those found in this study—primary closure in appro-
priately selected subjects appears to result in acceptable patient
n 68 68
outcomes with low rates of infection.6–9 Jenkinson et al8 in
Matched variables
a recent cohort study that used propensity score matching
Mean age, yr (SD)† 43.1 (19.7) 42.5 (18.4) 0.55
found very similar results to our study with significantly lower
Males, % 47 (69) 47 (69) 1.0
rates of infection when comparing matched subjects who
Grade, % 1.0
underwent either delayed or primary closure. Our study and
1 23 (34) 23 (34)
the recently published cohort study8 are 2 of the largest series
2 33 (48) 33 (48)
published to date with adjusted or matched comparison groups
3A 12 (18) 12 (18)
using modern fracture management approaches. These results
Fracture location, %‡ 1.0
may suggest that initial closure is not only acceptable, but may
Upper extremity 26 (38) 26 (38)
also be the preferred approach to the management of open
Femur 7 (10) 7 (10)
fracture wounds when possible. Lenarz et al (2010) has sug-
Tibia/fibula 35 (52) 35 (52)
gested using wound cultures to determine when wounds should
Comorbidities,% 0.009
be closed after open fracture. However, this large cohort study
0 34 (50) 17 (25)
of 422 subjects failed to show benefit in waiting for negative
1 18 (38) 30 (44)
wound cultures before wound closure.17
2 or more 16 (23) 21 (31)
This study has a number of strengths. As a prospective
Diabetes mellitus* 7 (10) 0 (0) ,0.001
study, we screened all patients admitted with open long bone
Current smokers at time of 41 (56) 32 (47) 0.17
injury, %
fractures. A standardized antibiotic regimen was followed and
Initial medical treatment 27 (43) 30 (44) 1.0
standardized forms were used at follow-up with a priori
elsewhere, % definitions of both infection and nonunion. In addition, we
Median time to antibiotic 3.25 (2.0–7.0) 3.25 (1.8–11.3) 0.99 achieved excellent (99%) follow-up (clinical and/or radio-
administration, hr (IQR) graphic) to fracture healing. In our matched-series analysis,
Median time to surgical 10.75 (7.8–16.1) 8.8 (6.2–11.3) 0.22 we drew subjects from the same center that had received
debridement, hr (IQR) similar perioperative and surgical care, completed the same
Received transfusions, % 22 (33) 16 (24) 0.34 data collection forms, and matched them on characteristics
Median hospital length of 5.5 (3.0–11.0) 6.5 (5.0–13.0) 0.61 known to impact the selected outcomes.
stay, d (IQR)§ The limitation of this study and the previous literature
Bolded P-value = significant at P , 0.05. on this topic are primarily related to nonrandomized study
*Counted in Comorbidities. designs and the inability to dictate the decision making of the
†P-value for age.
‡P-value for males, grade, fracture location, comorbidities, diabetes, current treating surgeon to perform primary closure. In all studies to
smokers and received transfusion. date, including the small randomized trial by Benson, the final
§P-value for median times: antibiotics, surgical debridement and hospital length decision to perform primary wound closure (or to include the
of stay.
patient in the study) was at the surgeons’ discretion.5–9 Even
with close adherence to a standardized management protocol,
including the use of postoperative antibiotics, surgeons must
still have the ability to determine what constitutes a “clean
than that reported in a recent systematic review of delayed wound” and is amenable to primary closure. Although our
closure after open fracture.1 Furthermore, in our matched study was population-based, we did exclude 28 fractures
analysis with subjects who underwent delayed closure in because of surgeons’ decision to not perform primary closure,
the same trauma center, we found both infection and non- so similar to other studies, we are somewhat limited in deter-
union to be significantly lower in the primary closure cohort. mining precise criteria for eligibility for primary wound clo-
Only 1 subject failed primary closure and returned to the sure. However, Moola et al (2014) and DeLong et al (1999)
operating room for further irrigation and debridement. This reported that they were able to use primary closure in 75%–
subject subsequently healed without further complications 88% of open fractures after implementing protocols that rec-
after the second debridement. ommended primary closure when possible, which suggests
The major concern with primary closure and the that this approach is possible for a majority of patients.6,9
recommendation to use delayed closure after repeat debride- Furthermore, although Lenarz et al17 (2010) suggested wait-
ment(s) has been based on early trauma evidence that primary ing for negative wound cultures before closure, this approach
closure increased the risk for deep infection with clostridium or did not lead to a decrease in infections.
other anaerobic organisms.2–4 However, with standardized anti- A second limitation of our work is that we included
biotic regimens and meticulous wound debridement, and appro- upper extremity fractures rather than focusing on lower
priate 48-hour wound check, these concerns may be less extremity fractures or tibial fractures, which have a higher
relevant in modern medicine. Furthermore, it is possible that risk of adverse outcomes such as infection when compared

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
with upper extremity.11 Upper extremity fractures accounted
for almost half of our enrolled subjects. With the low number
of adverse outcomes in our cohort, we were unable to do any
in-depth analysis of how fracture location may affect out-
comes when primary closure is used. Future work on primary
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.
In our matched-series analysis with similar subjects
who received delayed wound closure, we were also unable to
match on wound cleanliness between cohorts. We also did not
control for segment of bone fractured and note that there are
issues with the reliability of assigning Gustilo grades.13 Thus,
although primary closure was not commonly used in the ear-
lier cohort, our results should be interpreted cautiously as
clinical biases may exist between cohorts for which matching
did not control. Because our study was relatively small with
fewer than 100 subjects, we did not use propensity scoring,
but instead performed direct matching on a limited number of
variables to maximize the number of pairs achieved between
the 2 cohorts. However our results are consistent with the
current evidence on this topic.5–9
The surgeons involved in this study are experienced
trauma surgeons working at a high volume level 1 trauma
center with a large geographic catchment area. This may limit
the applicability of our findings to smaller centers with lower
patient volumes and less experienced surgeons who serve
a primarily urban-based populace, particularly because the
choice of primary or delayed wound closure was to some
degree at the surgeon’s discretion. However, our results are
similar to those recently reported by other high volume
trauma centers.8,9
Finally, we did not consider any economic impact of
primary versus delayed wound closure. We found that
subjects undergoing primary closure stayed a similar length
of time in hospital as those who underwent delayed wound
closure. We did not consider the cost of additional operating
room visits associated with a delayed wound closure
approach, which is an important economic consideration,
particularly when primary wound closure may lead to at least
equivocal and potentially superior patient outcomes relative
to delayed wound closure. Economic impact should also be
considered as an important outcome in any future trial.
CONCLUSIONS
In summary, the low rate of deep infection and
nonunion with primary wound closure for clean Gustilo
grade 3A or less fractures, lends credence to the practice of
primary closure in appropriately selected subjects. Although
126 | www.jorthotrauma.com
Copyright ! 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Orthop Trauma ! Volume 31, Number 3, March 2017
a definitive randomized trial that focuses on primary closure
in lower extremity fractures is likely required to answer this
question, this study supports the use of primary wound
closure in open fracture as it appears to reduce infection
and nonunion rates compared with delayed wound closure.
REFERENCES
1. Schenker ML, Yannascoli S, Baldwin KD, et al. Does timing to operative
debridement affect infectious complications in open long-bone fractures?
A systematic review. J Bone Joint Surg Am. 2012;94:1057–1064.
2. Hampton OP, Jr. Basic principles in management of open fractures.
JAMA. 1955;159:417–419.
3. Trueta J. The treatment of war fractures by the closed method. Proc R
Soc Med. 1939;33:65–74.
4. Trueta J. Reflections on the past and present treatment of war wounds
and fractures. Mil Med. 1976;141:255–258.
5. Benson DR, Riggins RS, Lawrence RM, et al. Treatment of open frac-
tures: a prospective study. J Trauma. 1983;23:25–30.
6. DeLong WJ, Born CT, Wei SY, et al. Aggressive treatment of 119 open
fracture wounds. J Trauma. 1999;46:1049–1054.
7. Hohmann E, Tetsworth K, Radziejowski MJ, et al. Comparison of
delayed and primary wound closure in the treatment of open tibial frac-
tures. Arch Orthop Traum Surg. 2007;127:131–136.
8. Jenkinson RJ, Kiss A, Johnson S, et al. Delayed wound closure in-
creases deep-infection rate associated with lower-grade open fractures:
a propensity-matched cohort study. J Bone Joint Surg Am. 2014;96:
380–386.
9. Moola FO, Carli A, Berry GK, et al. Attempting primary closure for all
open fractures: the effectiveness of an institutional protocol. Can J Surg.
2014;57:E82–E88.
10. Carsenti-Etesse H, Doyon F, Desplaces N, et al. Epidemiology of bac-
terial infection during management of open leg fractures. Eur J Clin
Microbiol Infect Dis. 1999;18:1999.
11. Weber D, Dulai SK, Bergman J, et al. Time to initial operative treatment
following open fracture does not impact development of deep infection:
a prospective cohort study of 736 subjects. J Orthop Trauma. 2014;28:
613–619.
12. Westgeest J, Weber D, Dulai SK, et al. Factors associated with develop-
ment of nonunion or delayed healing after an open long bone fracture:
a prospective cohort study of 736 subjects. J Orthop Trauma. 2016;30:
149–155.
13. Brumback RJ, Jones AL. Interobserver agreement in the classification
of open fractures of the tibia. The results of a survey of two hundred
and forty-five orthopaedic surgeons. J Bone Joint Surg Am. 1994;76:
1162–1166.
14. Gustilo RB, Mendoza RM, Williams DN. Problems in the management
of type III (severe) open fractures: a new classification of type III open
fractures. J Trauma. 1984;24:742–746.
15. Gustilo RB, Anderson JT. Prevention of infection in the treatment of one
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-
tion Compendium - 2007: Orthopaedic Trauma Association Classifica-
tion, Database and Outcomes Committee. J Orthop Trauma. 2007;21
Supplement 10 pp: S1–S163.
17. Lenarz CJ, Watson JT, Moed BR, et al. Timing of wound closure in open
fractures based on cultures obtained after debridement. J Bone Joint Surg
Am. 2010;92:1921–1926.
Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.