Rates of union and risk factors for continued nonunion following exchange nailing of tibial nonunion

Abstract

Introduction

The objective of this study was to evaluate the rate of nonunion repair success in tibia nonunions treated with exchange nailing. This retrospective cohort study was conducted across five academic Level 1 trauma centers and included 63 patients with tibia nonunions.

Materials and Methods

Patients who sustained a tibia fracture (AO/OTA 42) treated with intramedullary fixation that developed a nonunion and were subsequently treated with exchange nailing were retrospectively reviewed. The primary outcome measure was nonunion repair success based on osseous union. Additional analyses included union rate by AO/OTA classification, nonunion type, implant(s) used, graft used, time from initial procedure, and infection status.

Results

All patients sustained an AO/OTA type 42 fracture. Out of 63 patients, 47 tibias (75%) achieved osseous union after the index exchange nail procedure. The rates of nonunion revision success were similar across nonunion types and time from initial procedure until exchange nailing. There was no significant difference in union rate when infection was present. Complications included re-operation, readmission, infection, and implant failure.

Conclusions

This large, multicenter study with contemporary implants, instruments, and techniques for exchange nailing tibia nonunions demonstrates a nonunion repair success rate of 75%, consistent with the lower end of reported data in previous literature. No identifiable risk factors for failure of exchange nailing in tibia nonunions were found.

Level of Evidence

Level III Evidence

Introduction

Nonunion after intramedullary nailing of a tibia fracture is not uncommon, with rates ranging between 0 and 7% for closed tibia fractures [1–3] and between 8 and 32% for open tibia fractures, with even higher rates for Gustilo-Anderson class IIIB open tibia fractures [3, 4]. Fracture nonunions negatively affect both patients and healthcare systems. Patients with fracture nonunions have higher rates of pain, lower quality of life scores and are less likely to return to work [5–7]. Furthermore, tibia nonunions are associated with higher healthcare utilization and costs, higher indirect costs including productivity losses, and higher opioid use compared to tibia fractures that go on to union [8, 9].

High energy mechanism and open fractures are associated with increased rates of nonunion [10–13]. Tibia fractures may be more prone to nonunion than other bones due to their lack of muscle in areas within the circumferential soft tissue envelope [14]. One approach to manage tibia nonunions is reamed tibia exchange nailing. Exchange nailing includes removal of the original intramedullary nail, reaming the tibial canal, and inserting a new, larger diameter nail. Exchange nailing is postulated to treat nonunions through increased rigidity and strength, facilitating new vascular anastomoses, and through introduction of new biologically active tissue at the nonunion site [15].

The rate of union documented in the literature after tibia exchange nail varies; union rates range from 76 to 96% in aseptic diaphyseal tibia fractures treated with exchange nailing [16–20]. One study found a union rate of 63% in both aseptic and infected diaphyseal tibia fractures treated with an exchange nail [21]. Most of these retrospective studies consisted of surgeries performed twenty to forty years ago creating some question if the results will be similar in a more recent patient population with implants and instrumentation in more recent years. The purpose of our large multicenter study was to evaluate and update the rate of nonunion repair success following tibia exchange nailing with contemporary implants, instruments and techniques in order to inform treatment decisions. We hypothesized that there would be identifiable risk factors for failure of exchange nailing for tibia nonunions that would help guide decision making.

Materials and methods

Upon completion of Institutional Review Board approval, a retrospective review of diaphyseal tibia fractures (AO/OTA 42) that progressed to nonunion based on surgeon diagnosis after initial treatment with intramedullary nail were identified. We evaluated tibia nonunions treated at five Level 1 academic institutions. Radiographs were reviewed to confirm classification (hypertrophic, oligotrophic, atrophic) of nonunion diagnosis. Hypertrophic nonunions were identified as having abundant callus formation without bridging bone. Oligotrophic nonunions demonstrated incomplete callus formation. Atrophic nonunions were defined as having radiographically absent callus [22]. Concurrently, AO/OTA fracture classification and presence or absence of bony gap defect, defined as absence-in-continuity of bone, were also recorded.

Final inclusion criteria were defined as age 18 years or older, initial tibia fracture fixation with intramedullary nail, first nonunion procedure with exchange intramedullary nailing, and achieved union or a minimum of 3-month follow-up following the nonunion procedure. Patients treated with induced membrane at either initial fixation or nonunion surgery were excluded. All exchange nail procedures were performed by orthopaedic surgeons at the participating centers (n = 21 surgeons) between 2008 and 2020. Sixty-three patients met the above criteria for inclusion in the study.

Electronic medical records were reviewed to capture standard demographic data, tobacco use, laboratory values, open versus closed nature of initial injury, initial fracture fixation procedure data, nonunion surgical procedure information, and implants/biologics. Implants and biologics included intramedullary nail, plate, allograft or autograft (placed through a separate incision at the fracture site). Implant and biologic manufacturers were chosen by hospital and surgeon preference. Implant vendor choice was per surgeon preference. Dates of initial intramedullary nail and primary exchange nail were recorded. The length of time between the two procedures was subsequently documented and reported as a variable. Patient follow-up occurred per standard protocol at each participating Level 1 academic institution. Ultimate determination of success of exchange nail procedure was identified from radiographs obtained post exchange nail procedure as well as electronic documentation by individual orthopedic surgeons who managed each patient’s care. Standard post-operative complications including re-operation, re-admission for study injury, newly diagnosed post-operative infection, deep vein thrombosis, and implant failure were recorded.

We also evaluated all patients for evidence of infection based on laboratory markers and surgical cultures obtained at the time of nonunion surgery. Erythrocyte sedimentary rate (ESR) greater than 22 millimeters per hour (mm/hr) for males and 29 mm/hr for females and C reactive protein (CRP) greater than 3 milligrams per deciliter (mg/dL) were deemed abnormal [23]. Cultures taken at time of nonunion surgery were considered positive if they revealed a non-contaminant organism as determined by each hospital’s local laboratory. We then subdivided our patients into four categories: positive laboratory marker and culture, positive laboratory marker and negative culture, negative laboratory marker and positive culture, and negative laboratory marker and culture. The goal in doing so was to further stratify our patients and assess if statistically significant differences in outcomes occurred amongst these groups.

Statistical analyses were performed to determine the differences between the two groups, nonunion repair success versus recalcitrant nonunion, for each variable. All statistical analysis was conducted with SAS/STAT version 9.4 (Cary, NC). Chi-square, Fisher’s exact, and Kruskal-Wallis (nonparametric) tests were used in the analysis. An alpha level of < 0.05 was considered statistically significant.

Results

From a database of 706 long bone nonunions, we identified 63 tibias in 63 patients which met inclusion criteria and underwent exchange intramedullary nailing (Fig. 1). The time period for nonunion surgery for included patients ranged from January 2008 to February 2020. Average follow up was 10 months after exchange nail procedure. Patient demographics are summarized in Table 1, and some representative cases are shown in Fig. 2. Patients were predominantly male with a mean age of 38 years. 41% of patients smoked tobacco products. The majority of tibias were open injuries and were treated with both exchange intramedullary nailing and grafting at time of index nonunion surgery. Overall complication rate was 33% and included re-operation (n = 17, 27%), re-admission for the study injury (n = 10, 16%), new infection (n = 12, 19%), and implant failure (n = 5 patients). Most of the reoperations were either due to concern for infection, implant failure, or hardware removal. We only counted readmissions associated with the study injury. These mostly represented the patients who had reoperations that required an inpatient stay. Most of the infections were deep infections that required return to the operating room (n = 7/12, 58%). No significant differences in complication rates were found between patients who had successful nonunion repair (30%) and those who went on to recalcitrant nonunion (44%, p = 0.31).

Fig. 1.

Consort diagram for identification of tibia nonunions treated with exchange intramedullary nail (IMN). *Patients can have more than 1 exclusion criteria

Table 1.

Summary of demographics for 63 patients with tibia nonunion who underwent tibia exchange nail procedure

Age (Median, IQR)		38.5 (29, 48) years
Male Sex		52 (83%)
Diabetes		6 (9.5%)
Tobacco Use		26 (41%)
AO/OTA Classification	Type 42 A	11 (17%)
	Type 42B	28 (44%)
	Type 42 C	24 (38%)
Open		53 (84%)
Positive cultures at time of Nonunion Surgery		15 (24%)
Type of Nonunion	Hypertrophic	9 (14%)
	Oligotrophic	28 (44%)
	Atrophic	26 (41%)
Surgical Strategy	Exchange Nail	27 (43%)
	Exchange Nail + graft	36 (57%)

Note: Demographic Number of patients (n = 63)

Fig. 2.

Representative exchange nail cases. Tibia radiographs taken (A) and (B) at time of injury, (C) and (D) 6 months after initial intramedullary nail (IMN) fixation without evidence of union, and (E), (F), (G) 1 year after exchange IMN with fracture union

Forty-seven of 63 tibias (75%) were found to have nonunion repair success defined by bony union on radiographs after initial exchange nail procedure with an average time to union of 9.8 months. Gender, age, and tobacco use did not impact exchange nail success rate (Table 2). Nonunion type did not impact nonunion repair success rates after exchange nailing. Seven of 9 (78%) patients diagnosed with hypertrophic nonunion, 19 of 28 (68%) with oligotrophic nonunion, and 21 of 26 (81%) with atrophic nonunion went on to nonunion repair success after exchange tibia nail (p = 0.54). Further exploration to examine the impact of procedure type on nonunion type did not indicate differences in nonunion repair success (Table 3).

Table 2.

Rates of osseous union after tibia exchange nail (n = 63)

Total patients		Union	p value
		47 (75%)
Gender	Female (n = 11)	8 (73%)	0.99
Age (IQR)	38.5 years (29, 48)	38 (28, 47)	0.28
Current tobacco use	Yes (n = 26)	19 (73%)	0.99
AO/OTA	Type 42 A (n = 11)	8 (73%)
	Type 42B (n = 28)	20 (71%)	0.81
	Type 42 C (n = 24)	19 (79%)
Nonunion type	Hypertrophic (n = 9)	7 (78%)	0.54
	Oligotrophic (n = 28)	19 (68%)
	Atrophic (n = 26)	21 (81%)
Gap defect†	Yes (n = 14)	10 (71%)	0.74
Implant	Exchange nail (n = 27)	19 (70%)	0.51
	Exchange nail + graft (n = 36)	28 (78%)
Graft	Autograft alone (n = 33) Allograft alone (n = 0) Auto + allograft (n = 3)* No graft (n = 27)	27 (82%) n/a 1 (33%) 19 (70%)	0.13
Time of exchange nail from first procedure	Less than 6 months (n = 25)	21 (84%)	0.51
	6–12 months (n = 26)	18 (69%)
	Greater than 12 months (n = 9) Missing date of fixation (n = 3)	6 (67%) 2 (67%)
Closed vs. open	Open (n = 53)	41 (77%)	0.26
	Closed (n = 10)	6 (60%)
Positive cultures at time of nonunion surgery	Yes (n = 15)	9 (60%)	0.18
Culture outcomes	Positive (n = 15)	9 (60%)
	Negative (n = 25)	21 (84%)
	No culture (n = 23)	17 (74%)
Inflammatory labs / positive cultures**	Elevated labs, positive cultures (n = 7)	5 (71%)	0.25
	Elevated labs, negative cultures (n = 4)	3 (75%)
	Normal labs, positive cultures (n = 8)	4 (50%)
	Normal labs, negative cultures (n = 21) Labs and/or cultures not obtained (n = 23)	18 (86%) 17 (74%)

*Allograft: N = 1 BMP, N = 2 DBM

**Normal ESR: < 22 mm/hr (male) or < 29 (female); Normal CRP < 3 mg/dL

^† Gap defect at time of nonunion surgery

Table 3.

Procedure by nonunion type

	Implant			Graft
	Exchange nail (N = 27)	Exchange nail and graft (N = 36)	P	Autograft (N = 33)	Autograft and Allograft (N = 3)	None (N = 27)	P
Nonunion type
Atrophic nonunion	9 (33.3%)	17 (47.2%)	0.52	16 (48.4%)	1 (33.3%)	9 (33.3%)	0.61
Infected nonunion	6 (66.7%)	0		0	0	6 (66.7%)
Hypertrophic nonunion	4 (14.8%)	5 (13.8%)		4 (12.1%)	1 (33.3%)	4 (14.8%)
Infected nonunion	1 (25.0%)	1 (20.0%)		1 (25.0%)	0	1 (25.0%)
Oligotrophic nonunion	14 (51.8%)	14 (38.8%)		13 (39.3%)	1 (33.3%)	14 (51.8%)
Infected nonunion	6 (42.9%)	0		0	0	6 (42.9%)
Union
No	8 (29.6%)	8 (22.2%)	0.51	6 (18.1%)	2 (66.6%)	8 (29.6%)	0.13
Yes	19 (70.3%)	28 (77.7%)		27 (81.8%)	1 (33.3%)	19 (70.3%)
Infected Nonunion
Yes	13 (48.2%)	1 (2.8%)	0.0001	1 (3.0%)	0	13 (48.2%)	N/A
No	14 (51.8%)	35 (97.2%)		32 (97.0%)	3 (100%)	14 (51.8%)

The addition of graft to nonunion exchange nail procedure did not appear to improve the rate of healing. Of the 27 patients who received only an exchange nail, 19 patients (70%) united, while of 36 patients who received an exchange nail and graft, 28 patients (78%) united (p = 0.51). Nonunion repair success rates did not vary between graft type used including solely autograft or a combination of autograft and allograft, but we do not have sufficient numbers to determine whether or not this is significant. Thirty-three patients received autograft only, none received allograft only, and 3 patients received a combination of both auto and allograft for a total of N = 36 patients treated with graft. Most of those treated with grafts were open fractures (n = 31; 86%) and either atrophic or oligotrophic (n = 17 and n = 14, respectively; 86% collectively). Of those who received autograft, 4 patients received iliac crest autograft, 20 patients received local tibial autograft, 11 patients received autograft via reamer irrigator aspirator (RIA), and 1 received iliac crest and RIA. The presence of a segmental gap defect, defined as absence-in-continuity of bone, as interpreted by the primary orthopaedic surgeon on radiographs at time of nonunion surgery also did not affect nonunion repair success rates (p = 0.74). Fourteen patients were documented to have a segmental gap defect based on radiographic evidence at time of nonunion surgery, 10 of which (71%) went on to union after exchange nailing and half were treated with grafts.

The timing of exchange nail procedure did not significantly impact nonunion repair success rates. Out of 25 patients who underwent exchange nailing within six months of initial fracture fixation, 21 (84%) went on to nonunion repair success. For the 26 patients with nonunion exchange nail between 6 and 12 months from initial procedure, 18 (69%) achieved nonunion repair success. Six of 9 patients (67%) who underwent exchange tibia nailing greater than 12 months from initial surgery went on to nonunion repair success (p = 0.51).

There was no significant difference in nonunion repair success rates between open and closed fractures (p = 0.26). Patients with infection at time of nonunion surgery, defined as positive surgical cultures at time of nonunion surgery, had no significant difference in nonunion repair success rates after exchange nail procedure compared to patients without infection diagnosis, defined as no cultures taken or negative surgical cultures (p = 0.18). Of note, patients with positive surgical cultures were treated with culture-specific antibiotics under the guidance of their primary surgeon. The presence or absence of elevated inflammatory markers and positive surgical cultures also did not affect nonunion repair success rate after exchange nailing (p = 0.25). Out of 7 patients with both elevated inflammatory markers and positive surgical cultures, 5 patients went on to bony union after tibia exchange nail (71%). Furthermore, 3 of 4 patients with elevated inflammatory markers but negative cultures achieved osseous union (75%).

Discussion

This large, multicenter study utilizing contemporary implants, instruments, and techniques for exchange nailing of septic and aseptic tibia nonunions found a 25% failure rate. The rate of failure is disappointing, but it is consistent with prior literature that evaluated exchange nailing in both septic and aseptic tibia nounions [21]. AO/OTA classification of initial tibia fracture, type of nonunion, presence of a gap defect, implant used, presence or evidence of infection, and time to exchange nailing from first procedure did not appear to affect nonunion repair success. However, as discussed below, this study is likely underpowered to draw statistically significant conclusions.

We did not find a statistically significant difference in nonunion repair success when comparing patients with positive cultures to those with negative cultures or no cultures. We did see the highest rate of nonunion repair success among patients with a negative culture (84%), and the lowest rate of success among patients with a positive culture (60%). Patients without a culture had a success rate in the middle (74%). While our results are not statistically significant, the fact that patients with positive cultures had low rates of nonunion repair success is consistent with prior literature. Tsang et al. found that infection is a risk factor for failure of exchange nailing, where 65% of infected tibia nonunions failed to achieve osseous union after one exchange nail procedure, and 39% failed after a second exchange nail procedure [24]. Other studies have found that exchange nailing may introduce or increase the risk of surgical site infection and osteomyelitis [3, 13]. With adjunctive antibiotic treatment, exchange nailing of an infected tibial nonunion does not appear to have increased failure compared to exchange nailing for aseptic tibial nonunions.

Prior literature provides conflicting evidence regarding union rate after exchange nailing for different nonunion types. Multiple studies have demonstrated that reamed exchanged nailing is successful in atrophic and hypertrophic nonunions [25, 26]. Other studies suggest that oligotrophic and atrophic nonunion patterns increase the risk of exchange nail failure [21]. We found no convincing evidence of difference in nonunion repair success rates between hypertrophic, oligotrophic, and atrophic nonunions, suggesting that tibia exchange nails can be utilized in all types of nonunion patterns.

We found no evidence of difference in nonunion repair success rates between tibia nonunions treated with exchange nail alone versus exchange nail and grafting. Multiple prior studies have evaluated tibia exchange nails without the use of bone grafting [25, 26]. It is theorized that reaming alone produces intramedullary bone grafting that promotes bony union. These studies have found high rates of union for aseptic tibia nonunions treated with exchange nail alone [25, 26]. Another study has shown high rates of union in both septic and aseptic tibia nonunions treated with exchange nailing with or without bone grafting [21]. Importantly, percutaneous bone grafting is considered a safe and effective method to treat nonunions [27–30].

Our study found that timing of exchange nailing did not impact nonunion repair success rates. Of note, the SPRINT trial does suggest allowing a 6-month post-operative healing period after initial tibia intramedullary nailing prior to intervening for delayed healing or nonunion [14]. In this study, 28 patients underwent exchange nailing less than 6 months after initial tibia intramedullary nail.

Our patient population had an average time to union of 9.8 months after exchange nail procedure with an average follow up of 10 months after exchange nail procedure. Thus, the reported union rate of 75% is a likely a “floor” or minimum estimate of exchange nail success. It is possible that the actual success rate would be higher with longer patient follow up. There were twelve patients who were not united by final follow-up and did not have a reoperation. These patients may have been asymptomatic nonunions that the surgeons treated conservatively and plan to monitor over time, but the retrospective study design does not allow for documentation of reasons for decision making.

Strengths of this study include being a large, multicenter study with contemporary surgical implants and techniques. Furthermore, to our knowledge, this study is only the second to evaluate outcomes in exchange nailing for both septic and aseptic tibia nonunions. Given the ongoing debate regarding the use of exchange nailing in the setting of septic nonunions, we do believe this study provides an important contribution to the existing literature. A limitation of this study is that data was collected retrospectively rather than prospectively. While we did identify 63 patients who had undergone exchange nailing for tibia nonunions, this study is likely underpowered to draw statistically significant conclusions, particularly for subgroups of patients or a large number of variables. However, our retrospective multicenter analysis did allow us to collect a large cohort of patients, albeit not the largest documented tibia exchange nail cohort in both septic and aseptic tibia nonunions [21]. We also did not have a RUST score or CT to confirm nonunion diagnosis in this study. There were some patients who received an exchange nail procedure prior to 6 months. However, this represents the diagnostic and surgical practice patterns of a large group of orthopaedic surgeons at level 1 trauma centers across the country and therefore is likely generalizable. Another limitation is that complete data was not collected on amount of reaming performed and nail upsizing during the exchange nail procedure. Prior literature evaluating the success of exchange nailing has done so in the context of a defined reaming protocol with documentation of nail diameter for both the index and exchange nail [17, 19, 25]. Hierholzer et al. notes that ideally the exchange nail should be 2 mm larger than the original nail [17]. Increasing nail size is thought to facilitate the benefits of reaming, including debriding the intramedullary canal and improving local biology, while also mitigating the potential harms associated with over-reaming such as thermal necrosis [31, 32]. Increasing nail size also increases construct stiffness. Future studies should establish a defined reaming protocol and collect data on the amount of reaming performed as well as diameter of initial and exchange nails. We also did not collect data on fibular osteotomy. One multisite study on tibia nonunion did find that patients with fibular osteotomies proceeded to union faster than patients without fibular osteotomy [33].

A further limitation of this study is a lack of quantitative numerical data on bone defects. Previous studies suggest that exchange tibia nailing is most effective when bone defects are less than 50% of the cortical diameter or less than 2 cm total [3, 26]. Another study found that a bone gap greater than 5 mm was a risk factor for failure of tibia exchange nail [21]. However, our study only collected categorical data regarding a bone defect in the form of a yes/no format question. Future studies should obtain quantitative numerical information on bone defects and evaluate osseous healing in the context of this information. Overall, 22% of this study population had a segmental gap defect, and over half were treated with graft. While we do not have data regarding the indications for inclusion of grafting along with the exchange nail procedure, it is possible that this series represents more severe injuries as compared to similar studies in the literature reporting higher union rates.

Another limitation is lack of data collection on the use of antibiotic coated nails at time of exchange nail surgery. Prior studies have demonstrated that antibiotic coated intramedullary devices are a safe tool that can be used to improve outcomes in infected tibia nonunions [34–37].

Given that this study found no convincing evidence of difference in nonunion repair success based on any variable examined, we are left with the question of what factors do affect nonunion repair success in the setting of tibia nonunion exchange nailing. Furthermore, the appropriate selection criteria to identify patients who would benefit from tibia nonunion exchange nailing remains unclear.

In conclusion, this large, multicenter study with contemporary implants, instruments, and techniques for exchange nailing tibia nonunions demonstrates an osseous union rate of 75%, which is consistent with prior literature. Future studies should seek to examine why tibia nonunion exchange nailing continues to fail at relatively high rates, determine what steps can be done to improve results, and refine indications for the procedure.

Author contributions

JM, RS, and JH wrote the main manuscript text. ZY conducted the statistical analysis. All authors reviewed the manuscript.

Funding

This manuscript did not receive any funding.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Conflict of interest

Dr. Rivera reports participation on data safety monitoring board or advisory board: Adjudication committee for Bioventus research study (completed); leadership or fiduciary role in other board, society, committee or advocacy group, paid or unpaid: Committee of Board positions with AAOS, AAOSNow, ORS, and LLRS. Dr. Hsu reports consultancy for Globus Medical and Stryker, and personal fees from Smith & Nephew speakers’ bureau. Dr. Zura reports consulting fee for bioventus and stryker, payment for expert testimony for legal review, leadership or fiduciary role in other board, society, committee or advocacy group, paid or unpaid- kuntscher society. For the remaining authors, none were declared.