Abstract
Introduction:
Elderly patients are at risk of fracture nonunion, given the potential setting of osteopenia, poorer fracture biology, and comorbid medical conditions. Risk factors predicting fracture nonunion may compromise the success of fracture nonunion surgery. The purpose of this study was to investigate the effect of patient age on clinical and functional outcome following long bone fracture nonunion surgery.
Materials and Methods:
A retrospective analysis of prospectively collected data identified 288 patients (aged 18-91) who were indicated for long bone nonunion surgery. Two-hundred and seventy-two patients satisfied study inclusion criteria and analyses were performed comparing elderly patients aged ≥65 years (n = 48) with patients <65 years (n = 224) for postoperative wound complications, Short Musculoskeletal Functional Assessment (SMFA) functional status, healing, and surgical revision. Regression analyses were performed to look for associations between age, smoking status, and history of previous nonunion surgery with healing and functional outcome. Twelve-month follow-up was obtained on 91.5% (249 of 272) of patients.
Results:
Despite demographic differences in the aged population, including a predominance of medical comorbidities (P < .01) and osteopenia (P = .02), there was no statistical differences in the healing rate of elderly patients (95.8% vs 95.1%, P = .6) or time to union (6.2 ± 4.1 months vs. 7.2 ± 6.6, P = .3). Rates of postoperative wound complications and surgical revision did not statistically differ. Elderly patients reported similar levels of function up to 12 months after surgery. Regression analyses failed to show any significant association between age and final union or time to union. There was a strong positive association between smoking and history of previous nonunion surgery with time to union. Age was associated (positively) with 12-month SMFA activity score.
Conclusions:
Smoking and failure of previous surgical intervention were associated with nonunion surgery outcomes. Patient’s age at the time of surgery was not associated with achieving union. Advanced age was generally not associated with poorer nonunion surgery outcomes.
Introduction
The development of nonunion following bony fracture can be a significant challenge for both patients and treating orthopedic surgeons. In some cases, ununited fractures require multiple revisions and years of medical attention before bony healing is achieved. Several studies have elucidated risk factors associated with a failure to achieve union after fracture. 1 –5 Factors precluding healing can be grouped into those associated with mechanical instability or those associated with poor fracture biology. 1 An unstable fixation construct leads to potential shearing forces at the fracture site which limits both primary and secondary bone healing. 6 Biologic factors are sometimes implicated and include complex fracture patterns, bony comminution, interfragmentary gap, bone loss, poor soft tissue coverage, and vascular deficiency. 7 Patient comorbidities and social habits (particularly diabetes and nicotine use) are also strongly implicated. 5,8
Elderly patients may be at increased risk of development of nonunion, given that mechanical stability is more difficult to achieve (osteopenic bone), there is poorer overall fracture biology (diminished fracture vascularity and/or osteogenic potential), and comorbid conditions are more numerous and severe. The purpose of this study was 2-fold: (1) to determine whether increasing age was an independent risk factor for failure to achieve union after surgical repair of established long-bone nonunions and (2) to determine whether elderly patients with established nonunions who underwent surgical repair had different 1-year functional outcome scores when compared to their younger counterparts.
Materials and Methods
Between 2004 and 2012, 288 consecutive patients (age ranging 18-91 years) presenting to our tertiary care academic medical center with a long-bone nonunion were enrolled in a prospective research registry. The research protocol and all procedures followed were in accordance with the ethical standards determined by the institutional review board, and written informed consent was obtained from all study participants. Registry patients were treated by 4 trauma fellowship-trained orthopedic surgeons with experience in the care of these injuries. 2,9 –11
Basic demographic data, initial injury information, medical history, previous surgery, pain, and a baseline assessment of functional status using the Short Musculoskeletal Functional Assessment (SMFA) were recorded at presentation. Anatomic site and nonunion nature (atrophic/oligotrophic/hypertrophic) within each group were classified via the system described by Weber and Cech 12 and oligotrophic nonunions were combined with atrophic nonunions for the purposes of this study. In cases where the diagnosis was unclear on plain films, a computed tomography (CT) scan was obtained. After surgery, patients were scheduled for follow-up physical examinations at regular intervals (3, 6, and 12 months) at which time pain level and SMFA scores were reassessed. Radiographs were obtained at each visit to evaluate bony healing. Healing was determined radiographically using plain x-rays and clinically with physical examination. Healing was defined as the radiographic presence of bridging bone identified on 3 of 4 cortices without gross motion or tenderness at the site of nonunion. In situations where a determination could not be made, a CT scan was obtained to confirm union. 2,10,11,13
Intraoperative culture results and complications with wound infection identified during follow-up visits were recorded. Other follow-up complications, such as development of iliac crest graft site and surgical wound hematoma, wound dehiscence, wound breakdown, and malunion following revision, were also noted. In patients failing to reach union after primary revision, secondary revision surgery was documented. In situations where multiple revisions failed or in cases not amenable to revision, amputation may have been elected and such outcomes were recorded.
Surgical Intervention
All patients in this report were treated surgically for their fracture nonunion. The form of surgical intervention chosen was left to the discretion of the treating surgeon. The general principles applied during the assessment of injuries and decisions to perform surgery were similar for all surgeons. Patients without previous surgery underwent internal fixation; patients with previous surgery underwent revision internal fixation; and patients with suspected infection or those with limb deformity received external fixation revision. Eighty percent of patients received bone graft (either autogenous iliac crest or iliac crest aspirate), with or without bone morphogenetic protein bone morphogenetic protein 2 (BMP-2) or BMP-7 (recombinant human Osteogenic Protein 1, OP-1) adjunct.
Statistical Analysis
Patients failing to comply with the follow-up protocol were excluded. Patient’s age at the time of surgery was used to dichotomize the cohort into 2 groups in order to compare elderly (geriatric) patients (≥65 years of age) with nonelderly patients. Descriptive statistics were used to describe demographic characteristics, including gender, basal metabolic index (BMI), tobacco use, and medical comorbidities with particular attention paid to history of osteopenia or osteoporosis. Initial injuries were compared for open (vs closed) status, mechanism of injury (high energy vs low energy), and number of previous nonunion surgeries. Surgical outcomes including healing status, time to union, operating room (OR) cultures, wound complications, subsequent surgical revision or amputation, and 12-month Visual Analogue Scale (VAS) pain scores, and SMFA scores were compared. We performed a subgroup analysis of the most common elderly (femur) and nonelderly (tibia) nonunion site and evaluated differences in healing, time to union, reoperation, pain, and SMFA (all domains). In consideration of our dichotomization using age, a continuous variable (eg, a 64-year-old patient is not greatly dissimilar from a 65-year-old patient), we then performed an independent subgroup analysis (all fracture sites) comparing patients ≤40 years old with the elderly (≥65 years) group. In this additional analysis, we compared achieving union, time to union, reoperation, pain, and all indices of the SMFA.
Categorical variables were analyzed using chi-square tests while continuous data were analyzed using independent t-tests. Logistic regression analysis was used to determine whether age, gender, or known risk factors for nonunion (covariates: smoking status, initial open injury, BMI, and previous nonunion surgery) were independent predictors of failing to heal the nonunion. Linear regression analyses were used to determine the relationship of age with time to achieve union, 12-month VAS pain scores, and 12-month SMFA domain scores. All statistical analyses were performed with SPSS version 19.0 software (SPSS, Inc, Chicago, Illinois) and significance level set at P < .05.
Results
Sixteen patients (4 elderly and 12 nonelderly) did not comply with postoperative follow-up and were excluded from the analyses resulting in a total of 272 patients. Of the remaining 272 patients, 95.2% (259 of 272) achieved union at a mean of 7.1 ± 6.2 months (range 6 weeks-74 months). Fifty-six (20.6%) patients required a subsequent nonunion surgery to achieve healing. Thirteen (4.8%) patients developed a persistent nonunion despite revision attempts of which 3 followed-up for a year without additional intervention while 10 underwent subsequent revisions that also failed. Three (1.0%) patients, all with nonunion of the tibia and younger than 65 years of age, elected for amputation after undergoing several revision attempts.
The mean age of the elderly group was 73.1 ± 6.6 years old compared to 42.3 ± 12.6 years old for the nonelderly group (P < .01). Elderly patients statistically differed from nonelderly patients with regard to most of the studied demographic and injury variables except for BMI (Table 1). In all, 72% of the nonelderly group had undergone one or more previous nonunion surgeries while 54% of the elderly group had undergone previous nonunion surgery (mean = 1.4 [±1.5] vs 0.89 [±1.4], respectively; P = .02). There were no significant differences in the type of nonunion, atrophic nonunions being more common among both groups (elderly = 83.3%, nonelderly = 77.2%; P = .35; Table 2). There were differences in the anatomic sites of nonunion, as elderly patients were more likely to experience proximal long-bone (femur and humerus) nonunions, whereas the nonelderly group experienced a higher degree of tibial nonunions (P < .01).
Demographic Information, Patient Characteristics, and Baseline Injury Details.
Abbreviations: BMI, body mass index; SD, standard deviation; y/o, years old. Bold faced = reached statistical significance, as defined by the study to be p-value < 0.05.
Nonunion Type (Classified Via the System Devised by Weber and Cech 12 ) as Well as Anatomic Site of Fracture Nonunion.
Abbreviation: y/o, years old.
Autogenous bone grafting was performed in 80.1% (218 of 272) of all patients. There was no difference in rates of autogenous grafting between elderly and nonelderly patients (77.1% vs 80.8%; P = .55). A comparison of healing rates, complications, and outcomes is presented in Table 3. The elderly group healed at comparable rates (95.8% vs 95.1%, P = .59) with no statistical difference in healing time frame (6.2 ± 4.1 months vs 7.2 ± 6.6 months, P = .30) in comparison with the nonelderly group. The elderly group was just as likely to have positive OR cultures, wound complications, and reoperation as the younger group (Table 3).
Measures of Outcome Including Rates of Healing, Complications, and Standardized Functional Outcome Scores Using the Short Musculoskeletal Functional Assessment Questionnaire.
Abbreviations: SD, standard deviation; SMFA, Short Musculoskeletal Functional Assessment; OR, operating room; y/o, years old.
Two hundred and forty-nine (91.5%) patients completed a minimum of 1 year of follow-up with radiographs and SMFA survey evaluation performed at each visit. There were no differences in pain or any of the standardized SMFA score domains (Table 3). Subgroup analysis of femoral and tibial nonunions showed no differences between elderly and nonelderly groups with regard to healing, time to union, reoperation, pain, and SMFA (all domains; Table 4). Subgroup analysis comparing patients ≤40 years of age (n = 101) with the elderly patients (≥65 years, n = 48) revealed no statistical differences when comparing rates of healing (97.0% vs 95.8%, P = .71), time to union (7.9 [±8.8] vs 6.2 [±4.1] months, P = .2), reoperation (21.8% vs 12.5%, P = .18), VAS pain (2.58 [±2.3] vs 2.58 [±2.7], P = .99), and all indices of the 12-month standardized SMFA scores (Function P = .33, Bothersome P = .98, Activity P = .2, Emotion P = .83, Arm and Hand P = .06, and Mobility P = .48).
Subgroup Analysis of the Most Frequent Sites of Nonunion for Each Group (Elderly: Femur, Nonelderly: Tibia).a
Abbreviations: SD, standard deviation; SMFA, Short Musculoskeletal Functional Assessment; years old.
a Included in the analyses were a comparison of primary objective and subjective assessment measures, including standardized SMFA functional scores.
The association of age and statistically significant covariates with the measured outcomes is detailed in Table 5. Multivariate regression analysis revealed that age and other studied covariates were not predictors of achieving union after surgical repair. Time to union was not predicted by age, and the only covariates that were found to be predictive of this outcome were smoking status (P = .002) and previous nonunion surgery (P < .001). Twelve-month postoperative activity level (a domain of the SMFA) was the only outcome significantly associated with age, and regression analysis showed a 0.3 increase in standardized score for each annual increase in age (P < .03).
Multivariate Regression Analysis Including Age and Significant Covariate Associations With Time to Union, Pain, and 12-Month SMFA Functional Scores.a
Abbreviations: BMI, body mass index; CI, confidence interval; SMFA, Short Musculoskeletal Functional Assessment; VAS, Visual Analogue Scale; mos, months.
a Covariates included within the regression analysis were smoking, gender, BMI, initial open injuries, and number of previous nonunion surgeries.
Discussion
Repair of fracture long-bone nonunions is a difficult undertaking, given their compromised fracture-site biology. Host factors also play a role in inhibiting healing of established nonunions. Even after healing has occurred, host factors can negatively affect functional outcome. Age is a host factor that has not been studied extensively with regard to its role as a risk factor in inhibiting healing of established nonunions or its role in affecting short-term functional outcomes. It has been argued that increasing patient age may result in a change in the biomolecular environment, as progenitor mesenchymal cells of the elderly patients lack the osteogenic potential of younger patients. 14 –16 Regarding functional outcome, older patients may have more difficulty participating in rigorous physical therapy after trauma and surgery thereby potentially decreasing their overall mobility and functional outcome compared to their younger counterparts. 17 The finding of decreasing benefit from rehabilitation has been documented extensively in the arthroplasty literature. 18
After a review of the available literature, we believe that this is the first large series study focused on the effects of aging on operative success and functional outcomes after fracture nonunion surgery. A 2012 review article by Bishop and colleagues identified risk factors associated with compromised healing. 1 They highlighted the role of medical comorbidities (diabetes and vascular disease), age, gender, smoking, nonsteroidal anti-inflammatory drug use, and metabolic disease as patient-dependent risk factors for nonunion. Although the peer-reviewed literature supporting the link between smoking and nonunion is robust, 4,5,19 –21 an association between age and nonunion has not been as thoroughly studied. Several studies have documented the effect of age as a risk factor after operative and nonoperative management of clavicle fracture. 22,23 A small retrospective series by Green et al identified age as a risk factor for humeral fracture nonunion. 19 One study identified, by Brinker et al, and documented the evaluation of fracture nonunion surgery outcomes among a group tibial nonunion patients. 24 The study of Brinker and colleague included 23 patients with a mean age of 72 years (ranging 61-92) and is unique from ours in that they solely evaluated tibial nonunion treated patients with the Ilizarov method. Their study ultimately found that all patients completing treatment protocol (20 of 20) healed their tibial nonunion with a mean time of Ilizarov apparatus treatment of 9 months.
The goal of this study was to identify whether the elderly patient is at risk of failing nonunion repair surgery in the form of open reduction with internal fixation or dynamic external fixation. We found that the elderly patients were able to heal their injuries at similar rates and in a similar time frame as younger patients when treated by experienced surgeons. As evidenced within the demographics presented in this study, there are risk factors in the elderly patients that would seemingly put them in greater danger of failing nonunion surgery. In particular, the elderly patients had more than twice as many medical comorbidities. However, the elderly group also appeared to be protected from nicotine, experience less open injuries at baseline, and had fewer injuries caused by high-velocity mechanisms. Ultimately, the result was a group of patients who did just as well, if not marginally better, than their younger comparison counterparts.
There are several limitations to this study. In order to enhance the power of this study, we grouped fractures of all long-bone types together. Although we did identify trends toward increased prevalence of specific nonunion types in the nonelderly (tibial nonunion) group versus the elderly group (femoral nonunion), subgroup analyses of these anatomic sites yielded no differences in objective and subjective measures of outcome. There were 4 surgeons operating with slightly different preferences for their chosen methods of treatment. There was no standard protocol or algorithm employed by this study. Furthermore, we did not account for differences in treatment methods in our analyses. As noted, the majority of patients (80%) received autogenous graft or aspirate as part of their surgical intervention with no difference between elderly and nonelderly patients (77.1% vs 80.8%; P = .55), some receiving other adjuvants as well in a noncontrolled manner.
In conclusion, age alone should not be considered a contraindication for nonunion repair. Ultimately, we believe that these findings are encouraging for surgeons and their elderly patients who are considering surgical intervention for an established nonunion of a long bone. Nonunion surgery in the general population remains as a daunting challenge for even the most skilled of surgeons. In light of this study’s findings, it does not appear that the surgeon should consider advanced age as a unique and additional risk factor for the success of nonunion surgery. On the contrary, patients should be educated on the associations of current smoking and nonunion surgical failure with time to union. Although halting nicotine intake is difficult, it is a modifiable risk factor which appears important to the success of both young and old patients, alike. In contrast, failure of previous fracture nonunion surgery is not a modifiable risk factor. Certainly, this finding should caution surgeons to consider all management options in cases of recalcitrant nonunions, as the potential for serial surgical revision success appears to diminish.
Footnotes
Author’s Note
This study was presented, in part, at the 29th Annual Meeting of the Orthopaedic Trauma Association: Phoenix, AZ, October 9-12, 2013 and at the 44th Annual Meeting of the Eastern Orthopaedic Association: Miami, FL, October 30-November 2, 2013. This study received institutional review board approval.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.