Study Population
The patients in the current analysis constituted a subgroup of patients in
the LEAP study who were evaluated to assess the outcome of amputation or
reconstruction following a limb-threatening lower extremity
fracture4. For the
present analysis, we initially selected 197 patients with Gustilo type-IIIA
tibial fractures (fractures associated with major soft-tissue injuries [i.e.,
degloving or severe crush/avulsion injuries]), type-IIIB tibial fractures, and
type-IIIC tibial
fractures13, with
the exclusion of those who had coexisting limb-threatening foot, ankle, pilon,
or knee injuries. After the exclusion of patients who had segmental fractures
involving the proximal or distal part of the tibia (thirteen), those who had
less than two years of follow-up (twenty-two), and those who had delayed
amputation (six), 156 patients were available for study. One hundred and five
of these 156 patients underwent limb salvage, and the other fifty-one
underwent an early amputation and served as a control group. Of the 105
patients who underwent limb salvage, eighty-four had a Gustilo type-IIIB
injury, seventeen had a type-IIIA injury, and four had a type-IIIC injury.
Three of eighty-four type-IIIB injuries were treated with a fasciocutaneous
flap, whereas the remaining eighty-one were treated with a free muscle flap
(fifty) or a rotational muscle flap (thirty-one); hence, the statements in the
present report pertain to muscle flaps.
Of these 156 patients, 78% were evaluated by means of a telephone interview
at a minimum of eighty-four months after hospital discharge. Because the
seven-year follow-up was limited to a telephone interview, only questions that
were needed to compute the physical and psychosocial subscale scores were
asked. No orthopaedic or physical assessment was conducted at that time.
Characterizing the Patients and Injuries
Extensive baseline data on social and economic resources, demographic
characteristics, and health habits were collected and described
previously12. All
leg injuries were prospectively classified, at the time of admission and
soft-tissue coverage, with use of standard classification systems and
limb-salvage indices proposed in the
literature14-23.
Outcome Measures
The primary outcome measure used in the present analysis was the Sickness
Impact Profile
(SIP)24-28.
We also examined a number of important clinical outcomes, including (1) the
fracture-healing status at each follow-up time-point as assessed by the
treating surgeon (including L.X.W. and M.J.B.), who recorded the number of
weeks between admission and union, (2) the diagnosis of wound infection or
osteomyelitis requiring inpatient treatment, (3) the diagnosis of nonunion
requiring inpatient treatment, (4) the number of surgical interventions, (5)
the number of weeks to full weight-bearing, (6) the total number of days of
hospitalization, (7) the walking speed, (8) the percent impairment rating
based on range of motion according to the American Medical Association's
Guides to the Evaluation of Permanent
Impairment29,
and (9) the number of weeks between admission and return to work as reported
by the (previously employed) patient (see Appendix).
Data Analysis
The primary goal of the analysis was to examine the effect of clinical
variables that are hypothesized to predict functional and clinical outcomes in
the population of patients with a type-III open tibial diaphyseal fracture
(classification 42 according to the system of the Orthopaedic Trauma
Association14).
Major studied parameters were tabulated (see Appendix). In all cases, the
first analysis performed was to separate our study population by groups on the
basis of a single variable of interest and to compare the relevant outcomes
with use of simple bivariate tests (the chi-square test for proportions, the T
test for continuous variables, and analysis of variance for multiple
comparisons). In addition, relationships that were observed to be significant
at the bivariate level were also examined with use of multivariate regression
techniques, which can adjust for the effects of potential confounders.
Continuous variables were studied with use of multivariate regression
analysis, categorical variables were studied with use of logistic regression
analysis, and time to fracture-healing was studied with use of standard
survival analysis. Robust estimation techniques were used for all of the above
analyses4.
Specifically, we wanted to adjust for (1) the effect of sociodemographic
confounders, which have been previously shown to affect functional outcomes in
this population12,
and (2) the effect of injury characteristics and severity. Controlling for
injury characteristics and severity was particularly important in this
analysis because a primary goal of the study was to compare the outcomes of
patients in different treatment pathways. As it is likely that patients were
entered into these differing pathways because of different injury
characteristics and severity, we created models in which we adjusted for all
available injury descriptors. Because the results of the analyses presented in
this report were substantially unchanged by the addition of sociodemographic
and injury covariates, all of the significance values reported in this
manuscript are those from the simple bivariate significance tests.
In the present study, we identified treatment techniques and clinical
decisions that are controlled by the surgeon that could potentially modulate
the final outcome.
Specifically, it appears that the timing of wound débridement
(within six hours after the injury as compared with six to twenty-four hours
after the injury), the timing of soft-tissue coverage (three days or less
after the injury as compared with more than three days after the injury), and
the timing of bone-grafting procedures (less than three months after the
injury as compared with three months or more after the injury) did not impact
the infection or union rates and had no effect on functional outcome. These
findings are in contradistinction to what has been discussed in textbooks and
previous
reports8,9,30-43.
The multivariable analysis reported here is consistent with the report by
Harley et al.9 in
that there was no change in outcome or the percentage of patients having a
major complication when fractures that were débrided within six hours
after the injury were compared with those that were débrided between
six and twenty-four hours after the injury.
When there is associated periosteal stripping or exposed hardware and
simple coverage techniques are inadequate, muscle flap coverage is called for
because of its perceived benefits in terms of the enhancement of vascularity
and fracture-healing as well as the minimization of bacterial
growth32-35,41,42.
Although the timing of muscle flap wound coverage for the open tibial fracture
has been considered to be critical to a favorable clinical
outcome35-40,
no advantage was demonstrated among patients in whom this was accomplished
within the first seventy-two hours.
With respect to the data shown in Table
III concerning p-amp
scores4, these
findings support the strategy of an appropriately timed conversion to an
intramedullary nail for definitive fixation in instances in which external
fixation is initially used for so-called damage control or because of local
soft-tissue
considerations44-48.
The ideal time for bone-grafting in a patient with a high-energy tibial
shaft fracture is also an unsettled issue in the
literature43,49.
The analysis of the sixty-one patients in the current series who received a
bone graft failed to definitively clarify this issue, although those who
received a graft within three months after the injury trended toward better
outcomes at two years (in terms of both overall and physical SIP scores) than
did those who received a bone graft at or after three months (p <
0.01).
The limitations of the present study need to be considered. The method and
type of external fixation utilized across the centers was not delineated in
detail; thus, comparison between multiplanar and uniplanar fixation was not
possible. The follow-up evaluation at seven years was conducted by means of a
telephone interview with its associated
limitations50,51.
This follow-up evaluation was performed for 78% of the cohort, and the
inability to examine the patients or to assess radiographic joint changes
limited our ability to determine the factors that caused the deterioration of
the SIP scores over
time5. Although it
is acknowledged that some of the difference in the SIP scores between two and
seven years of follow-up can be ascribed to the mode of testing and, to some
extent, to the decay in follow-up, the aging of the
cohort52 and the
emergence of sequelae such as adjacent joint arthritis also probably
contributed to the differences that were seen.
In summary, surgeons caring for patients who have severe open tibial
diaphyseal fractures should consider that these patients may benefit from an
appropriately timed definitive stabilization of the fracture with an
intramedullary nail. Furthermore, if the surgeon is confronted with the need
to combine an external fixator as a definitive fixation with a muscle flap,
the patient and the treating team should consider that the outcomes of such
treatment are worse than those of amputation. Finally, we failed to observe
any advantages related to the timing of wound débridement (within six
hours as compared with six to twenty-four hours after the injury) or the
performance of early muscle flap wound coverage (within the first seventy-two
hours after the injury).
A table showing the details of the study parameters is available with the
electronic versions of this article, on our web site at jbjs.org (go to the
article citation and click on "Supplementary Material") and on our
quarterly CD-ROM (call our subscription department, at 781-449-9780, to order
the CD-ROM). ?
LEAP (Lower Extremity Assessment Project) Group: Michael J. Bosse, MD, and
James F. Kellam, MD, Department of Orthopaedic Surgery, Carolinas Medical
Center, Charlotte, NC; Ellen J. MacKenzie, PhD, Thomas G. Travison, PhD, and
Renan C. Castillo, MS, The Center for Injury Research and Policy, Johns
Hopkins University Bloomberg School of Hygiene and Public Health, Baltimore,
MD; Andrew R. Burgess, MD, and Andrew Pollack, MD, The R Adams Cowley Shock
Trauma Center, University of Maryland at Baltimore, Baltimore, MD; Lawrence X.
Webb, MD, Department of Orthopaedic Surgery, Wake Forest University Baptist
Medical Center, Winston-Salem, NC; Marc F. Swiontkowski, MD, and Doug Smith,
MD, Department of Orthopaedic Surgery, Harborview Medical Center, Seattle, WA;
Roy W. Sanders, MD, The Orthopaedic Trauma Service, Tampa General Hospital,
Tampa, FL; Alan L. Jones, MD, and Adam Starr, MD, Department of Orthopaedic
Surgery, University of Texas Southwestern Medical Center, Dallas, TX; Mark P.
McAndrew, Department of Orthopaedics and Rehabilitation, Vanderbilt University
School of Medicine, Nashville, TN; Brendon M. Patterson, MD, Department of
Orthopaedic Surgery, Cleveland MetroHealth Medical Center, Cleveland, OH; and
Melissa L. McCarthy, ScD, Johns Hopkins School of Medicine, Baltimore, MD.