Abstract
Background:
Knee stiffness is an important complication after periarticular fracture, but a systematic evaluation of risk factors for this complication and outcomes of treatment has not been undertaken, to our knowledge. The aims of this study were to evaluate risk factors for knee stiffness requiring manipulation after periarticular fracture and to document the clinical outcomes of the manipulation.
Methods:
This study was designed as a case-control study in which patients requiring manipulation under anesthesia after periarticular fracture were compared with those who did not require manipulation. Using billing data from a regional level-I trauma center, we identified twenty-four knees requiring manipulation for refractory stiffness over a six-year period. These were matched, on the basis of the AO/OTA classification, with forty-three control knees that did not develop stiffness requiring manipulation. Descriptive statistics were used for frequency and mean analysis.
Results:
Univariate analysis revealed that extensor mechanism disruption (chi square = 0.05), fasciotomy (chi square = 0.020), wounds requiring ongoing management and precluding knee motion (p = 0.001), and the need for more than two surgical procedures to achieve definitive fracture fixation and soft-tissue coverage (p = 0.003) all placed patients at increased risk for knee stiffness requiring manipulation. The mean improvement in knee motion following all procedures targeting knee stiffness was 62°. Mean final flexion was significantly less in the case group (107°) compared with the control group (124°; p = 0.01).
Conclusions:
To our knowledge, this is the first study to systematically evaluate the risk factors for knee stiffness after periarticular fracture and document the outcomes of manipulation under anesthesia. It demonstrates that injury characteristics that delay or prevent postoperative knee motion place patients at increased risk for refractory knee stiffness. Although knee motion remains compromised, late surgery aimed at improving knee motion leads to improvements in flexion.
Level of Evidence:
Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.
Loss of motion is a well-recognized and sometimes debilitating complication of knee surgery1-11. Operative procedures associated with knee stiffness include total knee arthroplasty, knee ligament reconstruction, and periarticular fracture surgery1-11. Severe cases of postoperative stiffness that do not respond to physical therapy can be treated with manipulation under anesthesia and other adjunctive procedures10-18.
Risk factors for knee stiffness, treatment options, and outcomes have all been well described in the arthroplasty and sports medicine literature. Risk factors for stiffness after total knee arthroplasty include diminished preoperative knee motion, previous surgery, depression, and poor pain tolerance1,8,19. In the sports medicine literature, acute reconstructive surgery, more extensive soft-tissue injury, and postoperative immobilization have all been associated with an increased risk of stiffness3,7,12. Technical errors by the surgeon and postoperative infection have been associated with increased rates of stiffness in both of these subsets of patients1,3,8,18. Depending on the severity and chronicity of the stiffness, treatment options generally include manipulation under anesthesia, arthroscopic lysis of adhesions, and more extensive open procedures. While postoperative outcomes can be difficult to interpret given the heterogeneity of the various pathological conditions and interventions, long-term function is often diminished.
In contrast to the relatively advanced understanding of knee stiffness after arthroplasty or ligament injury, stiffness after periarticular fracture is poorly understood. While knee stiffness has been reported as an important complication after fracture, a systematic evaluation of risk factors and outcomes has not been undertaken, to our knowledge5,9-11,16,17,20-25. The aims of this study were to evaluate risk factors for knee stiffness requiring manipulation after periarticular fracture and to document the clinical outcomes.
The study was designed as a matched case-control study with a ratio of 1:2 (two controls for each case). Billing records at a regional level-I trauma center were used to identify all patients who underwent manipulation under anesthesia (Current Procedural Terminology [CPT] code 27570) between October 2003 and October 2009. A total of 194 knees and 228 manipulations were identified. Although other procedures such as open or arthroscopic lysis of adhesions and quadricepsplasty are also employed to treat knee stiffness, they are always performed at our institution in conjunction with knee manipulation and were therefore also captured by our search.
A number of patients identified through the billing records had undergone knee ligament examination under anesthesia for determination of stability rather than manipulation for stiffness, and all of these patients were excluded. Patients with stiffness associated with arthroplasty or ligamentous knee injury were excluded as the pathology is distinct, and the risk factors and outcomes have already been described, in such cases. Patients requiring manipulation after below-the-knee amputation were also excluded. Because we used manipulation under anesthesia as a proxy for stiffness, any patients who had undergone manipulation for which knee stiffness was not the primary indication were also excluded. Such patients included many who had undergone knee manipulation when they returned to the operating room for treatment of other extremity fractures during their initial hospitalization or those who had undergone manipulation when the primary indication for surgery was wound management or treatment of infection. The goal of these manipulations was to prevent stiffness rather than to treat it, and we did not think these patients should be included. This left twenty-two patients with a total of twenty-four knees that we defined as stiff and that required at least one surgical procedure primarily to address loss of motion. These knees comprised the eligible case group for the study.
Using the orthopaedic trauma database at our institution, we attempted to match each case with two controls on the basis of age and injury constellation as classified with use of the AO/OTA system26. A 1:2 ratio was chosen because two cases for each control was the maximum number we were able to consistently identify in our database. The cases included four limbs that had sustained a combination of AO/OTA 33C and 41C-type injuries. Probably as a result of the severity and complexity of this injury constellation, we were able to match only four control limbs with this group. Similarly, one of the cases was a combination of AO/OTA 33A and 41C injuries, for which we also had difficulty finding a matching control. Ultimately, a single control with AO/OTA 33B and 41C injuries was identified for this case. Overall, the matching process yielded forty-three controls to compare with twenty-four cases.
Patient charts were reviewed to determine demographic data, injury mechanism, associated injuries, injury severity score (ISS), length of intensive care unit (ICU) stay, injury classification as defined by the AO/OTA and Gustilo27 systems, presence or absence of extensor mechanism injury, presence and duration of spanning external fixation, need for fasciotomy, number of surgical procedures, timing of manipulation, utilization of continuous passive motion, discharge to home or a rehabilitation facility, and knee motion before and after treatment. The presence of a wound requiring ongoing management in the form of serial irrigation and debridement, open wound care, or flap or skin graft coverage, and precluding early knee motion, was recorded with use of this stated definition. The injury characteristics as defined by the AO/OTA and Gustilo classification systems are summarized in the Appendix. Beyond this, we were unable to quantify the extent of articular injury or the presence of meniscal injury.
Data Analysis
Data analysis was performed with use of SPSS software (version 17; Chicago, Illinois). Univariate analysis was performed with use of means and frequencies for the descriptive statistics. Group t tests were used to compare continuous variables, and chi-square tests were used to compare nonparametric variables. Backward conditional logistic regression was performed with use of the variables found to be significant in the univariate analysis, with p < 0.05 as the criterion for entry into and p > 0.10 as the criterion for removal from the equation. We did not link the case control as triplets in our logistic regression when we used backward stepwise conditional elimination.
Source of Funding
There was no external funding source for this investigation.
Demographic Data
The basic demographic and diagnostic characteristics of the study groups are summarized in Table I.
Manipulation
Manipulation under general anesthesia was carried out at a mean of 124 days (range, twenty-eight to 622 days) after definitive fixation. In fourteen cases, the manipulation was carried out in isolation. Ten cases involved additional procedures such as lysis of adhesions, excision of heterotopic ossification, quadricepsplasty, and removal of implants. After an initial manipulation, four patients required additional procedures to treat refractory stiffness. This included one patient treated with repeat manipulation under anesthesia; one treated with excision of heterotopic bone and repeat manipulation; one treated with repeat manipulation, lysis of adhesions, and implant removal; and one treated with manipulation, lysis of adhesions, and quadricepsplasty. This resulted in a mean of 1.2 procedures per stiff knee specifically aimed at restoring motion. Preoperative, intraoperative, and final knee motion data were extracted from the medical record. The mean duration of follow-up was 364 days (range, seventy-six to 1923 days) in the control group and 365 days (range, eighty-seven to 830 days) in the case group.
Range of Motion
In the case group, mean extension and flexion averaged 3° (range, 0° to 20°) and 45° (range, 15° to 75°), respectively, before the manipulation and 1° (range, 0° to 5°) and 120° (range, 85° to 140°) immediately following the initial manipulation. A total of twenty-eight procedures aimed at improving motion were performed in the twenty-four case knees, and mean extension and flexion averaged 4° (range, −5° to 15°) and 107° (range, 30° to 150°) at the time of final follow-up. In three cases, 90° of knee flexion was not achieved. At the time of final follow-up in the control group, mean extension and flexion averaged 2° (range, 0° to 10°) and 124° (range, 95° to 150°), respectively. Final flexion in the controls was significantly greater than that in the cases (p = 0.01). Final extension did not differ significantly between the groups (p = 0.12). The average improvement in knee flexion following all procedures targeting stiffness was 62°.
Effect of Extensor Mechanism Injury
An extensor mechanism disruption was present in four (17%) of the twenty-four cases and in one (2%) of the forty-three controls. Extensor mechanism injury significantly increased the risk of knee stiffness requiring manipulation (chi square = 0.05) (Table I).
Effect of Fasciotomy
Fasciotomy was performed in five (21%) of the twenty-four cases and in one (2%) of the forty-three controls. The risk of knee stiffness requiring manipulation was significantly higher in the patients who required fasciotomy (chi square = 0.020) (Table I).
Effect of Wound Management
Ongoing wound management in the form of dressing changes, serial debridement, delayed closure, or eventual flap coverage was necessary in ten (42%) of the twenty-four cases and in three (7%) of the forty-three controls. By definition, all patients who underwent fasciotomy also required ongoing wound management. The need for ongoing wound management significantly increased the risk of knee stiffness requiring manipulation (p = 0.001) (Table I).
Effect of Spanning External Fixator
Fifteen (63%) of the twenty-four cases and seventeen (40%) of the forty-three controls were treated with a spanning external fixator. This difference was not significant (chi square = 0.21). Analysis of the knees placed in a spanning external fixator revealed that the duration of fixation averaged twenty-one days (range, four to 139 days) for the cases and eight days (range, two to nineteen days) for the controls. This difference between groups did not reach significance (p = 0.16) (Table I); however, there was a clinically important difference in the number of days of treatment with a spanning external fixator between the two groups. Sixty-five percent of the controls (eleven of seventeen) were treated with the spanning external fixator for eight days or less and 29% (five of seventeen), for three days or less; only 27% of the cases (four of fifteen) were treated with the spanning external fixator for less than eight days. There was an outlier in the case group—a patient who was treated with the fixator for 139 days—and this contributed to the lack of significance. If this case was removed, the difference between the two groups would demonstrate significance (p < 0.05).
Effect of Number of Surgical Procedures Required
An average of three surgical procedures (range, one to fourteen) were required to achieve definitive fracture fixation and soft-tissue coverage in the case group compared with two surgical procedures (range, one to five) in the controls; this difference was significant (p = 0.021) (Table I). When this parameter was evaluated as a continuous variable, it was not found to predict the need for manipulation. Patients were also characterized as requiring two or fewer procedures (typically initial spanning external fixation followed by definitive internal fixation) compared with more than two procedures to achieve definitive fracture fixation and soft-tissue coverage. Twelve (50%) of the twenty-four cases required more than two procedures on the involved limb, including all patients who were initially treated with fasciotomy. Seven (16%) of the forty-three controls required more than two procedures. This difference was significant, with more than two procedures being associated with an increased risk of manipulation under anesthesia (p = 0.003).
A backward conditional logistic regression model was conducted with the criterion for variable entry set at p < 0.05 and that for variable exclusion from the equation set at p > 0.10. Backward conditional logistic regression was performed with use of the variables identified as significant in univariate analysis. The model revealed that the factors that were significant predictors of the need for manipulation were wound management (odds ratio, 10.8 [95% confidence interval, 2.5, 46.6]) and extensor mechanism injury (odds ratio, 11.1 [95% confidence interval, 1.1, 114.5]) (Table II).
Other Factors Not Affecting Risk of Stiffness Requiring Manipulation
Analysis of age, sex, side, and mechanism of injury for the cases versus the controls did not show any to be associated with an increased risk of stiffness. Discharge home or to a rehabilitation facility, use or no use of postoperative continuous passive motion, presence or absence of a head injury, presence or absence of an open fracture, ISS, and length of ICU stay were also not associated with increased risk (Table I).
To our knowledge, this is the first study to systematically valuate the risk factors for knee stiffness after periarticular fracture and to document the outcomes of manipulation under anesthesia. Understanding these factors may help surgeons to better identify patients at risk for substantial stiffness, provide targeted preoperative and postoperative counseling, and potentially treat high-risk patients with more intensive postoperative rehabilitation protocols. Furthermore, we have provided new data on the efficacy of knee manipulation under anesthesia in the setting of postfracture stiffness. This information should prove helpful to surgeons who confront this difficult problem.
Extensor mechanism disruption, fasciotomy, wounds requiring ongoing management and precluding knee motion, and the need for more than two surgical procedures to achieve definitive fracture fixation and soft-tissue coverage place patients at increased risk for stiffness requiring manipulation. One plausible mechanism by which these risk factors may all influence stiffness is by precluding early aggressive knee motion. Although the duration of spanning external fixation was not a significant predictor of stiffness, our data suggest that it is a clinically important variable for identifying patients at risk.
It is difficult to compare the risk factors identified in this study with those associated with stiffness after arthroplasty, as total knee replacement is typically not performed to treat acute injuries in young patients and consequently many baseline patient characteristics are quite different. Preoperative stiffness and previous surgical procedures are less relevant to patients with an acute traumatic injury, many of whom have no history of knee pathology. While depression is certainly prevalent in an orthopaedic trauma population and poor coping mechanisms affect outcomes after fracture, we did not assess these factors in our patient population28,29. Ligamentous reconstructions about the knee, however, are frequently performed to treat acute traumatic injuries and the risk factors for stiffness after such procedures are similar to those that we have identified, with increasingly severe soft-tissue injury and a lack of early knee motion portending a poorer prognosis3,12.
Factors not associated with an increased risk of stiffness requiring manipulation are age, sex, injury mechanism, discharge disposition (home versus rehabilitation), use or no use of postoperative continuous passive motion, presence or absence of a head injury, presence or absence of open fracture, ISS, and length of ICU stay. This suggests that the skeletal and soft-tissue characteristics of the injury itself play a more important role than the overall status of the patient, at least in terms of the return of knee motion. While the role of continuous passive motion after total knee arthroplasty continues to be debated, we found no evidence that it was of benefit after complex periarticular fracture.
One previous study that evaluated predictors of outcome after floating knee injuries did identify the presence of a Gustilo type-3 open fracture as a risk factor for stiffness; however, all of the fractures in that study were treated with either delayed skin-grafting or flap coverage, presumably compromising postoperative rehabilitation22. At our institution, many type-3A open fracture wounds are closed primarily after aggressive irrigation and debridement, potentially facilitating earlier knee motion. Our study suggests that it is not the presence of the open fracture specifically but rather the implications for postoperative rehabilitation that are prognostically important in terms of final knee motion.
There are several limitations to our study. Knee motion measurements were extracted from the medical record and were not standardized, likely adversely affecting accuracy. We also matched cases with controls partially on the basis of the AO/OTA classification. Therefore, we did not evaluate the effect of the injury site on the development of stiffness. Authors of previous publications have reported high rates of stiffness after so-called floating knee injuries, and the high percentage of floating knees among our cases would certainly support this notion, although our study was not specifically designed to answer this particular question22,30. We were also unable to find two controls for every case, so five case limbs had only one control each. Therefore, we performed unconditional logistic regression rather than conditional logistic regression analysis.
There are additional limitations related to the retrospective nature of the study. We chose the need for manipulation under anesthesia as a proxy for refractory stiffness but acknowledge that there were likely patients whose outcomes were adversely affected by stiffness but for whom manipulation was not performed. Participation in physical therapy may also play an important role in the return of knee motion. Although our institutional protocol involves routine physical therapy after periarticular fracture with early and aggressive motion exercises that continue on an outpatient basis, our study design precluded us from systematically evaluating patient participation. We recognize that all patients who underwent fasciotomy were also characterized as requiring ongoing wound management, and we were unable to determine whether the fasciotomy itself confers an increased risk or if it is just another manifestation of a wound. Similarly, both fasciotomy and wound management were correlated with the need for an increased number of surgical procedures. Nevertheless, we utilized a large database of periarticular fractures in order to carry out this case-control study, examining many variables and evaluating a large number of patients. Although stiffness after knee arthroplasty and ligamentous reconstruction has received substantial attention, this is the first study to focus specifically on knee stiffness after fracture, to our knowledge.
The present study demonstrates that injury characteristics that delay or prevent postoperative knee motion place trauma patients at increased risk for refractory knee stiffness requiring manipulation. Although knee motion remains compromised, late manipulation aimed at improving knee motion does lead to improvements in knee flexion.
A table showing the distribution of injuries according to the AO/OTA and Gustilo classifications is available with the online version of this article as a data supplement at jbjs.org.
Bong
MR;
Di Cesare
PE. Stiffness after total knee arthroplasty. J Am Acad Orthop Surg.
2004 May-Jun;12(
3):164-71.
DeHaven
KE;
Cosgarea
AJ;
Sebastianelli
WJ. Arthrofibrosis of the knee following ligament surgery. Instr Course Lect.
2003;52:369-81.[PubMed]
Magit
D;
Wolff
A;
Sutton
K;
Medvecky
MJ. Arthrofibrosis of the knee. J Am Acad Orthop Surg.
2007 Nov;15(
11):682-94.
Nelson
CL;
Kim
J;
Lotke
PA. Stiffness after total knee arthroplasty. J Bone Joint Surg Am.
2005;87
Suppl 1(Pt 2):264-70.[PubMed][CrossRef]
Papagelopoulos
PJ;
Partsinevelos
AA;
Themistocleous
GS;
Mavrogenis
AF;
Korres
DS;
Soucacos
PN. Complications after tibia plateau fracture surgery. Injury.
2006 Jun;37(
6):475-84. .[CrossRef]
Ries
MD;
Badalamente
M. Arthrofibrosis after total knee arthroplasty. Clin Orthop Relat Res.
2000:177-83.
Robertson
GA;
Coleman
SG;
Keating
JF. Knee stiffness following anterior cruciate ligament reconstruction: the incidence and associated factors of knee stiffness following anterior cruciate ligament reconstruction. Knee.
2009 Aug;16(
4):245-7. .[CrossRef]
Schiavone Panni
A;
Cerciello
S;
Vasso
M;
Tartarone
M. Stiffness in total knee arthroplasty. J Orthop Traumatol.
2009 Sep;10(
3):111-8. .[CrossRef]
Vander Have
KL;
Ganley
TJ;
Kocher
MS;
Price
CT;
Herrera-Soto
JA. Arthrofibrosis after surgical fixation of tibial eminence fractures in children and adolescents. Am J Sports Med.
2010 Feb;38(
2):298-301. .[CrossRef]
Warner
JJ. The Judet quadricepsplasty for management of severe posttraumatic extension contracture of the knee. A report of a bilateral case and review of the literature. Clin Orthop Relat Res.
1990;(
256):169-73.
Lee
DH;
Kim
TH;
Jung
SJ;
Cha
EJ;
Bin
SI. Modified Judet quadricepsplasty and Ilizarov frame application for stiff knee after femur fractures. J Orthop Trauma.
2010 Nov;24(
11):709-15.[CrossRef]
Shelbourne
KD;
Patel
DV. Treatment of limited motion after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc.
1999;7(
2):85-92.[PubMed][CrossRef]
Alici
T;
Buluç
L;
Tosun
B;
Sarlak
AY. Modified Judet’s quadricepsplasty for loss of knee flexion. Knee.
2006 Aug;13(
4):280-3. .[CrossRef]
Fox
JL;
Poss
R. The role of manipulation following total knee replacement. J Bone Joint Surg Am.
1981 Mar;63(
3):357-62.
Esler
CN;
Lock
K;
Harper
WM;
Gregg
PJ. Manipulation of total knee replacements. Is the flexion gained retained?J Bone Joint Surg Br.
1999 Jan;81(
1):27-9.[CrossRef]
Jovanović
S;
Orlić
D;
Wertheimer
B;
Zelić
Z;
Has
B. Quadricepsplasty after war fractures. Mil Med.
2000 Apr;165(
4):263-7.
Massè
A;
Biasibetti
A;
Demangos
J;
Dutto
E;
Pazzano
S;
Gallinaro
P. The Judet quadricepsplasty: long-term outcome of 21 cases. J Trauma.
2006 Aug;61(
2):358-62.[CrossRef]
Scuderi
GR. The stiff total knee arthroplasty: causality and solution. J Arthroplasty.
2005 Jun;20(
4 Suppl 2):23-6.[CrossRef]
Gandhi
R;
de Beer
J;
Leone
J;
Petruccelli
D;
Winemaker
M;
Adili
A. Predictive risk factors for stiff knees in total knee arthroplasty. J Arthroplasty.
2006 Jan;21(
1):46-52.[CrossRef]
Charnley
J. Knee movement following fractures of the femoral shaft. J Bone Joint Surg Am.
1947 Jul;29(
3):679-86.
Hahn
SB;
Choi
YR;
Kang
HJ;
Lee
SH. Prognostic factors and long-term outcomes following a modified Thompson’s quadricepsplasty for severely stiff knees. J Bone Joint Surg Br.
2010 Feb;92(
2):217-21.
Hee
HT;
Wong
HP;
Low
YP;
Myers
L. Predictors of outcome of floating knee injuries in adults: 89 patients followed for 2-12 years. Acta Orthop Scand.
2001 Aug;72(
4):385-94.[CrossRef]
Hutson
JJ
Jr;
Zych
GA. Treatment of comminuted intraarticular distal femur fractures with limited internal and external tensioned wire fixation. J Orthop Trauma.
2000 Aug;14(
6):405-13.[CrossRef]
Mariani
PP;
Del Signore
S;
Perugia
L. Early development of patella infera after knee fractures. Knee Surg Sports Traumatol Arthrosc.
1994;2(
3):166-9.[PubMed][CrossRef]
Mitchell
SE;
Keating
JF;
Robinson
CM. The treatment of open femoral fractures with bone loss. J Bone Joint Surg Br.
2010 Dec;92(
12):1678-84.[CrossRef]
Marsh
JL;
Slongo
TF;
Agel
J;
Broderick
JS;
Creevey
W;
DeCoster
TA;
Prokuski
L;
Sirkin
MS;
Ziran
B;
Henley
B;
Audigé
L. Fracture and dislocation classification compendium - 2007: Orthopaedic Trauma Association classification, database and outcomes committee. J Orthop Trauma.
2007 Nov-Dec;21(
10 Suppl):S1-133.[CrossRef]
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 Jun;58(
4):453-8.
Crichlow
RJ;
Andres
PL;
Morrison
SM;
Haley
SM;
Vrahas
MS. Depression in orthopaedic trauma patients. Prevalence and severity. J Bone Joint Surg Am.
2006 Sep;88(
9):1927-33.[CrossRef]
Starr
AJ. Fracture repair: successful advances, persistent problems, and the psychological burden of trauma. J Bone Joint Surg Am.
2008;90
Suppl 1:132-7.[PubMed][CrossRef]
Lundy
DW;
Johnson
KD. “Floating knee” injuries: ipsilateral fractures of the femur and tibia. J Am Acad Orthop Surg.
2001 Jul-Aug;9(
4):238-45.
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.