Tibial shaft fractures are common injuries that have a significant incidence of nonunion. The reported nonunion rate of tibial shaft fractures ranges from 5% to 15%.1–3 Tibial non-union leads to prolonged recovery for the patient and adds significant medical and societal costs. Costs were 2.3 times greater in patients who developed tibial non-union compared with those who healed primarily.4 In patients with high-energy open tibial fractures, average healing time was 10 months, and 37% of patients never returned to work.5
Owing to prolonged and costly recovery, surgeons have recommended early secondary interventions such as exchange intramedullary nailing (IMN) and bone grafting in patients having at-risk tibial fractures to avoid nonunion and prolonged recovery.6,7 Evidence has demonstrated that a significant percentage of tibial fractures heal between 6 and 12 months after injury without secondary interventions.8 Therefore, clinical recommendations regarding the treatment of delayed or un-healed tibial shaft fractures include waiting at least 6 months prior to secondary intervention.9 A treatment dilemma exists framed by patients having tibial shaft fractures that develop nonunion leading to secondary surgery and patients who heal on a delayed basis without further intervention. Generating evidence that could accurately identify which patients would benefit from early secondary surgery and which patients would predictably heal and avoid unnecessary surgery would be a significant advance.
Patients with tibial fractures are invariably closely followed postoperatively. During office visits, changes in clinical examinations, including patient-reported symptoms, gait, and radiographic fracture healing, are routinely queried. In this study, the authors investigated the utility of a simple office-based clinical tool called the tibial fracture healing score (TFHS), which uses basic components of a standard postoperative clinical examination and radiographic analysis to predict the need for secondary intervention in patients with tibial shaft fractures. They hypothesized that the TFHS can predict the need for secondary intervention to obtain union in patients with tibial shaft fractures compared with clinical examination and radiographic assessment alone.
Materials and Methods
This prospective study received institutional review board approval. Patients 18 years and older who had tibial shaft fractures (AO/OTA 41A, 42A-C, and 43A)10 treated with reamed, locked IMN at a Level I trauma center from 2013 to 2017 were included. During the study period, 380 acute tibial shaft fractures were treated with IMN (Figure 1). Patients with operative ipsilateral lower extremity injuries, pathologic fractures, age younger than 18 years, planned nonunion surgery secondary to critical size defect, nonunion surgery performed prior to 6 months after index IMN at surgeon discretion, infected nonunion, inability to consent, and anticipated follow-up noncompliance were excluded (n=239). Patients with nondisplaced fracture extension into the knee and/or ankle were included as long as a separate open approach for articular fracture fixation was unnecessary. Patient demographics and injury characteristics were obtained via patient interview and electronic chart review. Clinical and radiographic data were collected at routine follow-up intervals: 2 and 6 weeks and 3, 6, and 12 months. Eight fellowship-trained orthopedic trauma surgeons participated in the study. Based on previous investigations evaluating tibial nonunion, the diagnosis of nonunion was defined by persistent pain and lack of radiographic healing for greater than 6 months after the index procedure.6,11,12
CONSORT flow diagram of study participation. Abbreviation: IMR, intramedullary rod.
An office-based clinical score accounting for patient subjective information, physical function, and physical examination was calculated by surgeon assessment of the following 3 parameters at routine office visits:
Subjective: Patients who have mild or no pain or a significant decrease in pain compared with their prior visit based on treating physician judgment receive a score of 1. The key component to score a point was a clearly documented decrease in pain compared with the previous visit by the treating physician (none/mild/decreased=1, no change/increased=0).
Function: Patients who walk unassisted (or back to their previous level of ambulating if they had a walker or other equipment prior to their injury) without a limp or minimal limp (limp does not slow down a normal cadence of walking) and can control and balance a single-leg stance in a steady manner score 1 point (minimal limp/able to perform a single-leg stance=1, significant limp/unable to perform single-leg stance=0).
Physical examination: Patients who have minimal or no pain on deep manipulation of the fracture site receive 1 point. Deep manipulation of the fracture was performed by applying a bending moment to the fracture in both the coronal and sagittal planes. Hardware-associated pain remote to the fracture site does not affect this score (no/minimal pain with deep manipulation=1, pain with manipulation=0).
The sum of the subjective, function, and physical examination scores, ranging between 0 and 3, was defined as the clinical healing score. Radiographic healing was assessed by the radiographic union scale in tibial fractures (RUST) score.13 The RUST scores assign an integer (1 to 3) individually to each of the 4 cortices visualized on the anteroposterior and lateral radiographs, for an overall score ranging from 4 to 12. Subsequently, the authors calculated an adjusted RUST (aRUST) score by dividing the RUST score by 4 to equally weight the clinical and radiographic components to calculate the TFHS. The TFHS is the sum of the clinical score (0 to 3) and the aRUST score (1 to 3), yielding a score ranging from 1 to 6. The authors hypothesized that the TFHS would more accurately predict tibial shaft nonunion compared with either the RUST score or the clinical score in isolation at the 3-month office visit. The RUST scores, clinical scores, and TFHS were measured at the 3-month follow-up visit.
Receiver operating characteristic (ROC) curves were constructed to evaluate the accuracy of the RUST score, the clinical score, and the TFHS in predicting nonunion. Patient demographics and injury characteristics were compared using the Mann–Whitney U and Fisher's exact tests between the union and non-union groups. For all analyses, 2-tailed P values were used and deemed significant if less than .05.
Of 141 patients enrolled, 54 patients were excluded due to insufficient follow-up before diagnosis of union or nonunion. Eighty-seven patients were included in the analyses. The nonunion rate was 11%. Demographic and injury characteristics were compared between patients with union and nonunion (Table 1). There was no difference in age, sex, compartment syndrome, tobacco use, chronic disease, American Society of Anesthesiologists (ASA) score, AO/OTA classification, and flap coverage between the groups. The authors found a significant difference in non-union rates in open fractures (P=.01) and a trend toward significance in high-energy mechanism injuries (P=.05) (Table 1).
Patient Demographics by Outcome
The median time to diagnosis of union was 22 weeks (mean, 25±12 weeks). The median time to diagnosis of nonunion was 32 weeks (mean, 32±6 weeks). The mean clinical examination score was 2.6 (median, 3±0.7) for union and 0.6 (median, 0.6±0.7) for nonunion. The mean aRUST score was 2 (median, 2±0.4; RUST score, 8) for union and 1.3 (median, 1.25±0.28; RUST score, 5.2) for nonunion. The mean TFHS was 4.6 (median, 4.75±0.80) for union and 1.9 (median, 1.75±0.82) for nonunion.
The analysis of ROC curve demonstrated that a RUST score of 5 or less yielded 97% sensitivity (95% confidence interval [CI], 0.93–1.00), 70% specificity (95% CI, 0.40–1.00), area under the curve (AUC) 0.931, 78% positive predictive value (PPV), and 98% negative predictive value (NPV) in predicting the need for secondary surgery to obtain union (Figure 2). A RUST score of 8 or greater yielded sensitivity of 65% (95% CI, 1.00–1.00), specificity of 100% (95% CI, 0.55–0.75), 27% PPV, and 100% NPV in predicting need for secondary surgery to obtain union (Table 2). RUST scores of 6 or 7 yielded 84% sensitivity (95% CI, 0.75–0.92), 80% specificity (95% CI, 0.54–1.00), 20% PPV, and 97% NPV. Thirty percent (3 of 10) of patients in the nonunion group and 33% (25 of 77) of patients in the union group had RUST scores of 6 and 7 (Table 2). Clinical scores of less than 2 yielded 91% sensitivity (95% CI, 0.84–0.96), 90% specificity (95% CI, 0.70–1.00), AUC 0.956, 56% PPV, and 99% NPV in predicting the need for secondary surgery to obtain union (Table 2). When aRUST scores were added to clinical scores, a TFHS score of less than 3 moderately increased sensitivity and NPV (96% sensitivity [95% CI, 0.92–1.00], 90% specificity [95% CI, 0.70–1.00], AUC 0.986, 75% PPV, and 99% NPV) in predicting the need for secondary surgery to obtain union (Figure 2, Table 2).
Three separate receiver operating curves demonstrating improved sensitivity and specificity with tibial fracture healing score (TFHS) in predicting nonunion when compared with clinical score and radiographic healing alone. Abbreviations: aRUST, adjusted radiographic union scale in tibial fractures; RUST, radiographic union scale in tibial fractures.
Sensitivity, Specificity, Positive Predictive Value, and Negative Predictive Value of Assessments
The nonunion rate in this series of 87 patients whose tibial fractures were treated with reamed, locked IMN was 11% overall (4% for closed fractures and 24% for open fractures) and is comparable with rates reported in previous studies.1–3,8,14–17 A trend toward statistical significance in high-energy mechanism injury and delayed healing is consistent with previous observations.14,18 The current authors hypothesized that the TFHS, informed by typical information collected during a standard postsurgical office visit, would more accurately identify patients who heal successfully or progress to nonunion compared with either a radiographic score or a clinical score in isolation. In the current study, patients with tibial shaft fractures who had at least 2 of 3 favorable key clinical examination findings (symptoms/pain, function, and physical examination of fracture site) at 3 months were likely to heal without a secondary procedure (99% NPV). The addition of radiographic assessment to clinical examination did not improve NPV, likely due to the fact that radiographic healing lags behind clinical improvement in fracture healing. However, radiographic healing was more evident in the union group (mean RUST score, 8 vs 5.2); therefore, when radiographic scores were added to the clinical examination score for TFHS, the PPV improved from 56% to 75%. The authors showed modest improvements in identifying patients at risk of nonunion with increases in sensitivity and PPV compared with RUST score and clinical score.
Tibial fracture healing has been extensively studied, with particular focus on early identification of patients at risk for nonunion. However, it remains difficult for clinicians to accurately stratify nonunion risk. Surgeons predicted tibial nonunion with a sensitivity and specificity of only 74% and 62%, respectively, based on a retrospective analysis of a composite panel of patient information, including age, medical comorbidities, smoking status, and radiographs collected 3 months after injury.19 Radiographic-based analyses have also proved to poorly predict tibial fracture healing, but it is difficult to interpret the literature secondary to discrepancies in radiographic definitions of fracture healing.13,20–25 Investigators have standardized radiographic indices of tibial fracture healing by using the RUST score methods. This has improved interobserver reliability and proved to be better than both the surgeon's overall impression and assessment of the degree of cortical bridging.24,25 A recent study described that any cortical bridging (minimum RUST score, 6) by 4 months postoperatively reliably predicted tibial shaft fracture healing treated with IMN with an accuracy of 99%.14 Similar findings were observed in the current study, as a RUST score of 5 or less modestly predicted nonunion at 3 months.
O'Halloran et al14 identified risk factors for tibial fracture nonunion based on medical comorbidities, fracture characteristics, and radiographs at the time of fracture fixation to create a nonunion risk determination score.15 They found that low-energy mechanism of injury pattern, female sex, closed fracture, and 100% cortical contact reduced nonunion rates. On the other hand, decrease in cortical contact, increase in ASA score, chronic disease, compartment syndrome, and requiring flap coverage increased the risk of nonunion. The nonunion risk determination score provides a prediction model to allow surgeons to identify high-risk patients at the time of index fracture fixation surgery. Collectively, these studies demonstrate an extensive body of literature dedicated to risk stratifying tibial nonunion. Noticeably absent in both the RUST score and nonunion risk determination score methods is information pertaining to function assessment and physical examination. The impact of positive or negative findings on physical examination, in the presence of inconclusive or contradictory radiographic findings, is unclear and presents a treatment dilemma to the treating physician. The ability to objectively establish the PPV or NPV of these clinical findings would likely improve decision making and patient care. Improvements in predictions of the trajectory of tibial fracture healing with the TFHS likely reflect supplementing the score with meaningful physical findings, including pain and function.
Interestingly, ROC analyses demonstrated that a RUST score of 8 or greater at 3 months will likely mean healing without secondary intervention, whereas a RUST score of 5 or less will likely mean that a secondary intervention will be required to achieve union. Uncertainties for fracture healing exist in patients with RUST scores of 6 or 7 (84% sensitivity, 80% specificity). Inevitably, decision making for secondary surgery to achieve union is more difficult in such patients as opposed to when there is an obvious lack of healing or clear evidence of healing on radiographs. More than 30% (n=28) of patients in this study had RUST scores of 6 or 7. Ninety-two percent (23 of 25) of patients with a RUST score of 6 or 7 who healed had a clinical score of 2 or greater. On the other hand, all 3 patients in the nonunion group had a clinical score of 0 or 1. This finding may suggest that the TFHS may be particularly useful for identifying patients with radiographic delayed healing who will ultimately require a second procedure.
This study had several limitations. The sample was small and there was loss to follow-up, which may have induced selection bias and compromised the validity of the results. Patients who had non-union surgery prior to 6 months based on surgeon discretion were excluded. These patients could have affected the results because they would likely have had low TFHS. However, it is unlikely that they would have had numerical effects on the results because there were only 6 of them. Inherent bias of surgeon-dependent and patient-dependent clinical examination score measurements may have affected the results. The 3 parameters in the clinical score all include interpretation by the surgeon and are openly subjected to surgeon bias. For example, the subjective component that grades improvements in pain is clearly open to interpretation by the surgeon. The physical function test is less affected by surgeon interpretation because it is fairly objective to determine whether a patient can perform single-leg stance successfully. However, these biases were weighed against practicality. All 3 components of the clinical score can be readily tested and recorded within minutes and are common metrics measured in the outpatient practice setting.
Strengths of this study included prospectively collected data with no deviation from typical clinical practice. Furthermore, the study used RUST score for radiographic healing to avoid previously reported limitations in the reliability of assessment of radiographic fracture healing.
To the authors' knowledge, this study is the first to report a tibial fracture healing index with physical examination components. Stratifying healing or nonhealing trajectories will allow surgeons to expedite necessary nonunion interventions and avoid costs of unnecessary surgery. These results suggest that clinical examination and the TFHS are accurate in predicting successful tibial fracture healing and identifying patients at risk of nonunion. Patients with a clinical examination score of less than 2 and a TFHS of less than 3 at the 3-month postoperative visit should be followed closely, and this may mitigate delay in surgical intervention. The TFHS needs to be prospectively validated in a larger sample. It would also be of interest to apply the nonunion risk determination score criteria to later time points to see how they compare with the TFHS in a larger sample.
- Audigé L, Griffin D, Bhandari M, Kellam J, Rüedi TP. Path analysis of factors for delayed healing and nonunion in 416 operatively treated tibial shaft fractures. Clin Orthop Relat Res. 2005;438(438):221–232. doi:10.1097/01.blo.0000163836.66906.74 [CrossRef] PMID:16131895
- Zura R, Xiong Z, Einhorn T, et al. Epidemiology of fracture nonunion in 18 human bones. JAMA Surg. 2016;151(11):e162775. doi:10.1001/jamasurg.2016.2775 [CrossRef] PMID:27603155
- Gaston P, Will E, Elton RA, McQueen MM, Court-Brown CM. Fractures of the tibia: can their outcome be predicted?J Bone Joint Surg Br. 1999;81(1):71–76. doi:10.1302/0301-620X.81B1.0810071 [CrossRef] PMID:10068007
- Antonova E, Le TK, Burge R, Mershon J. Tibia shaft fractures: costly burden of nonunions. BMC Musculoskelet Disord. 2013;14(1):42. doi:10.1186/1471-2474-14-42 [CrossRef] PMID:23351958
- Saddawi-Konefka D, Kim HM, Chung KC. A systematic review of outcomes and complications of reconstruction and amputation for type IIIB and IIIC fractures of the tibia. Plast Reconstr Surg. 2008;122(6):1796–1805. doi:10.1097/PRS.0b013e31818d69c3 [CrossRef] PMID:19050533
- Court-Brown CM, Keating JF, Christie J, Mc-Queen MM. Exchange intramedullary nailing: its use in aseptic tibial nonunion. J Bone Joint Surg Br. 1995;77(3):407–411. doi:10.1302/0301-620X.77B3.7744925 [CrossRef] PMID:7744925
- Zelle BA, Gruen GS, Klatt B, Haemmerle MJ, Rosenblum WJ, Prayson MJ. Exchange reamed nailing for aseptic nonunion of the tibia. J Trauma. 2004;57(5):1053–1059. doi:10.1097/01.TA.0000100380.50031.DC [CrossRef] PMID:15580032
- Bhandari M, Guyatt G, Tornetta P III, et al. Study to Prospectively Evaluate Reamed Intramedullary Nails in Patients with Tibial Fractures (SPRINT) Investigators. Randomized trial of reamed and unreamed intramedullary nailing of tibial shaft fractures. J Bone Joint Surg Am. 2008;90(12):2567–2578. doi:10.2106/JBJS.G.01694 [CrossRef] PMID:19047701
- Schemitsch EH, Bhandari M, Guyatt G, et al. Study to Prospectively Evaluate Reamed Intramedullary Nails in Patients with Tibial Fractures (SPRINT) Investigators. Prognostic factors for predicting outcomes after intramedullary nailing of the tibia. J Bone Joint Surg Am. 2012;94(19):1786–1793. doi:10.2106/JBJS.J.01418 [CrossRef] PMID:23032589
- Meinberg EG, Agel J, Roberts CS, Karam MD, Kellam JF. Fracture and dislocation classification compendium: 2018. J Orthop Trauma. 2018;32(1)(suppl 1):S1–S170. doi:10.1097/BOT.0000000000001063 [CrossRef] PMID:29256945
- Brinker MR, O'Connor DP. Exchange nailing of ununited fractures. J Bone Joint Surg Am. 2007;89(1):177–188. doi:10.2106/JBJS.F.00742 [CrossRef] PMID:17200326
- Oh JK, Bae JH, Oh CW, Biswal S, Hur CR. Treatment of femoral and tibial diaphyseal nonunions using reamed intramedullary nailing without bone graft. Injury. 2008;39(8):952–959.
- Whelan DB, Bhandari M, Stephen D, et al. Development of the radiographic union score for tibial fractures for the assessment of tibial fracture healing after intramedullary fixation. J Trauma. 2010;68(3):629–632. doi:10.1097/TA.0b013e3181a7c16d [CrossRef] PMID:19996801
- O'Halloran K, Coale M, Costales T, et al. Will my tibial fracture heal? Predicting non-union at the time of definitive fixation based on commonly available variables. Clin Orthop Relat Res. 2016;474(6):1385–1395. doi:10.1007/s11999-016-4821-4 [CrossRef] PMID:27125823
- Lack WD, Starman JS, Seymour R, et al. Any cortical bridging predicts healing of tibial shaft fractures. J Bone Joint Surg Am. 2014;96(13):1066–1072. doi:10.2106/JBJS.M.00385 [CrossRef] PMID:24990971
- Sarmiento A. On the behavior of closed tibial fractures: clinical/radiological correlations. J Orthop Trauma. 2000;14(3):199–205. doi:10.1097/00005131-200003000-00010 [CrossRef] PMID:10791672
- Finkemeier CG, Schmidt AH, Kyle RF, Templeman DC, Varecka TFA. A prospective, randomized study of intramedullary nails inserted with and without reaming for the treatment of open and closed fractures of the tibial shaft. J Orthop Trauma. 2000;14(3):187–193. doi:10.1097/00005131-200003000-00007 [CrossRef] PMID:10791670
- Sanders R, Jersinovich I, Anglen J, Di-Pasquale T, Herscovici D Jr, . The treatment of open tibial shaft fractures using an inter-locked intramedullary nail without reaming. J Orthop Trauma. 1994;8(6):504–510. doi:10.1097/00005131-199412000-00008 [CrossRef] PMID:7869165
- Yang JS, Otero J, McAndrew CM, Ricci WM, Gardner MJ. Can tibial nonunion be predicted at 3 months after intramedullary nailing?J Orthop Trauma. 2013;27(11):599–603. doi:10.1097/BOT.0b013e31828f5821 [CrossRef] PMID:23481919
- Hammer RR, Hammerby S, Lindholm B. Accuracy of radiologic assessment of tibial shaft fracture union in humans. Clin Orthop Relat Res. 1985;(199):233–238. PMID:4042484
- Dijkman BG, Sprague S, Schemitsch EH, Bhandari M. When is a fracture healed? Radiographic and clinical criteria revisited. J Orthop Trauma. 2010;24(suppl 1):S76–S80. doi:10.1097/BOT.0b013e3181ca3f97 [CrossRef]
- McClelland D, Thomas PB, Bancroft G, Moorcraft CI. Fracture healing assessment comparing stiffness measurements using radiographs. Clin Orthop Relat Res. 2007;457(457):214–219. PMID:17159575
- Morshed S, Corrales L, Genant H, Miclau T III, . Outcome assessment in clinical trials of fracture-healing. J Bone Joint Surg Am. 2008;(suppl 1):62–67. doi:10.2106/JBJS.G.01556 [CrossRef]
- Whelan DB, Bhandari M, McKee MD, et al. Interobserver and intraobserver variation in the assessment of the healing of tibial fractures after intramedullary fixation. J Bone Joint Surg Br. 2002;84(1):15–18. doi:10.1302/0301-620X.84B1.0840015 [CrossRef] PMID:11837825
- Kooistra BW, Dijkman BG, Busse JW, Sprague S, Schemitsch EH, Bhandari M. The radiographic union scale in tibial fractures: reliability and validity. J Orthop Trauma. 2010;24(suppl 1):S81–S86. doi:10.1097/BOT.0b013e3181ca3fd1 [CrossRef] PMID:20182243
Patient Demographics by Outcome
|Characteristic||Nonunion (n=10)||Union (n=77)||P|
|Age, mean (range), y||36 (18–66)||36 (18–73)||.493|
|Male, No.||8 (80.0%)||45 (58.4%)||.304|
|Time to surgery, mean (range), h||10.5 (1.0–17.0)||12 (2–384)||.223|
|Open, No.||8 (80.0%)||26 (33.8%)||.012|
| Distal||7 (70.0%)||37 (48.1%)|
| Mid||1 (10.0%)||29 (37.7%)|
| Segmental||0 (0.0%)||3 (3.9%)|
| Proximal||2 (20.0%)||8 (10.4%)|
|Compartment syndrome, No.||1 (10.0%)||5 (6.6%)a||.535|
|High energy, No.||10 (100%)||54 (70.1%)||.057|
|Fibular fracture, No.||9 (90.0%)||64 (83.1%)||1.0|
|Former tobacco user, No.||6 (60.0%)||35 (46.1%)a||.509|
|Chronic disease, No.b||0 (0.0%)||9 (11.7%)||.590|
|Flap coverage, No.||1 (10.0%)||2 (2.6%)||.310|
|ASA score, No.||.959|
| 1||2 (20.0%)||17 (22.1%)|
| 2||5 (50.0%)||35 (45.5%)|
| 3||3 (30.0%)||18 (23.4%)|
| 4||0 (0.0%)||7 (9.1%)|
Sensitivity, Specificity, Positive Predictive Value, and Negative Predictive Value of Assessments
|RUST score ≤5||97%||70%||78%||98%|
|RUST score 6 or 7||84%||80%||20%||97%|
|RUST score ≥8||65%||100%||20%||100%|
|Clinical score <2||91%||90%||56%||99%|