Non-unions after intramedullary nailing of fractures of the humeral diaphysis

Summary

Aim of the work: The aim of this work is to evaluate the results of the treatment of the non-unions of hume­ral diaphysis after primary intramedullary nailing.

Material and method: From January 2010 to Ja­nuary 2017, 209 patients with fractures of the hume­ral diaphysis with an intramedullary nail were opera­ted at the Trauma Hospital in Brno. The group involved retrospectively evaluated patients with a nonunion of humeral diaphysis originated after primary treatment of fracture with an intramedullary nail These non­unions were treated by nail extraction and stabilization by autocompression or LCP. The shortest follow-up period of the patients after nonunion surgery was at least 12 months. Evaluated criteria of the group of non-unions included primary fracture types, types of nonunions, Constant shoulder score, disabilities of the arm, shoulder and hand (DASH) score, bone healing time after revision surgery and possible complications.

Results: 14 patients were included in the group of patients with nonunion following primary intramedullary osteosynthesis of the humeral diaphysis fracture addressed by nail extraction and splint stabilization.

In all patients, the nail was primarily inserted in the antegrade procedure. The mean follow-up after the revision surgery was 26 months. Eight cases were hypertrophic, two in oligotrophic and in four patients arthophic. The healing of the nonunion occurred in all operated patients in an average period of 4.1 months. The DASH score was 13 points at the end of the follow-up of the group of revision surgeries. Constant shoulder score was 77 points on average. After the revision surgery, 3 patients had transient paresis of the radial nerve. In two cases, nerve function was fully restored. In one case, distal forearm and hand hyperaesthesia persist.

Conclusion: Fractures of type A1 and B1 according to the AO classification appear to be predisposed for the possible development of nonunion. The solution of each nonunion should be approached individually with good preoperative planning.

Keywords:

Intramedullary nail – humeral diaphysis – nonunion

INTRODUCTION

Fractures of humeral diaphysis can be treated conservatively or surgically. Conservative treatment is still the method of choice for most closed fractures of the humeral diaphysis. Spiral fractures of the proximal and medial diaphysis of the humerus are at greater risk of developing a nonunion and are therefore indicated for surgical treatment [24, 30]. Surgical treatment includes splint osteosynthesis, intramedullary nail or external fixa­tor. Excellent results of tibia and femur fracture nailing offered the use of this surgical method for fractures of the humerus too. This surgical method has developed for humerus more in the last two decades. One advantage of nailing is the low invasiveness and lower incidence of infectious complications compared to splint osteosynthesis [11, 16]. For many authors, the use of the nail on the humerus is still perceived controversially, and some surgeons are reluctant to humerus nailing and prefer splint osteosynthesis [24].

The nonunions after conservative treatment of the humeral diaphysis fracture occur from 2 % to 20 % and from 8 % to 12 % [10, 29] after surgical treatment. The nonunion on the humerus is defined as a fracture without evident healing 24 weeks after the injury [10]. Brinker defines the nonunion as a state of fracture healing where, in the opinion of the attending surgeon, there is no chance of healing without further surgical intervention [5]. The solution to the nonunions of humeral diaphysis by removing the nail, pre-drilling the medullary cavity and replacing the nail with a larger diameter nail that works well on the lower limb does not show such good results due to different biomechanics on the upper limb [4]. Better results are obtained by conversion to splint osteosynthesis [3, 15]. Risk factors associated with the development of nonunion include age, smoking and alcoholism, open fracture, high-energy injury, infection and type of fracture [2, 17, 27].

The aim of the work is to retrospectively evaluate the results of the treatment of the nonunions of humeral diaphysis after primary intramedullary nailing by nail removal, nonunion revision and splint osteosynthesis.

MATERIAL AND METHODOLOGY

Patient group

From January 2010 to January 2017, 209 patients with fractures of the humeral diaphysis with an intramedullary nail were operated at the Trauma Hospital in Brno. The authors retrospectively evaluated the clinical results of patients with the nonunion of humeral diaphysis after primary stabilization of the fracture by intramedullary nail, who were treated by nail extraction and revision of the nonunion with plate fixation.

The initial criterion for the inclusion of patients in the follow-up was the fracture of the humerus. The group did not include patients with bilateral fracture, pathological fracture or patients primarily treated at another workplace, because it was not possible to trace the type of fracture. The classification of the fractures initially treated was performed according to the AO/OTA classification [21]. All nonunions were treated by nail extraction and stabilization by autocompression or LCP. In the case of arthophic nonunion, resection of sclerotic margins, decortication and bone graft application were performed. The shortest follow-up period of the patients after nonunion surgery was 12 months. Evaluated criteria included primary fracture types, types of nonunions, Constant shoulder score, disabilities of the arm, shoulder and hand (DASH) score [13], bone healing time and possible complications after revision surgery.

Surgical technique to address humeral diaphysis nonunion

The nail removal was performed through the original scars after previous extraction of the proximal and distal securing screw. In case of nail or locking screws breakage, the access was extended as needed, most often to Henry’s anterolateral, eventually a single incision was performed followed by trepanation of the medullar cavity and broken screws were removed. It was easier to remove the broken nail from the nonunion. N. radialis in the distal third of the humerus in the intermuscular septum between m. brachialis and m. brachioradialis was prepared and protected throughout the operation for anterolateral access.

If atrophic nonunion was found, the sclerotic margins of the nonunion were resected into bleeding and vital tissue (paprika sign). The intramedullary canal was excochlea­ted and treated with drills. This was followed by correction of the axis and rotational deviations and dressing of the bone margins so as to be able to remove the nonunion. Compression by a tension screw was performed in the indicated cases for oblique nonunions. Subsequently, the nonunion was stabilized with a 4.5 mm plate (3.5 mm for gracile skeleton) applied laterally with compression. For good bone quality, we used a conventional autocompression plate (DCP). The use of angularly stable plate (LCP) was indicated for osteoporotic bone. We fixed the plate with at least three bolts above and three bolts under the nonunion, introduced bicortically. Spongioplasty used for the atrophic nonunions was of autologous type from the ala ossis illium. In two cases we resected the atrophic nonunion and performed a shorte­ning with compression on the DCP plate. In both of these patients, the defect occurring after resection had up to 2 cm in length.

Hypertrophic nonunion were treated by removing the original, often failed nail, creating a plate bed, and application of the DCP or LCP plate. In recent years, even for hypertrophic nonunions, we have included nonunion drilling or chipping techniques with an oscillating saw [20]. We do not use bone grafts for hypertrophic nonunion.

RESULTS

In 18 cases (8.6 %), the nonunion occurred after primary intramedullary nailing. In all these cases the nail was initially introduced in the antegrade manner. In eight ca­ses, the nail broke and in seven cases, the locking screws broke and the nail migrated proximally.

Surgery was recommended for all patients with verified nonunion. Four patients refused surgery most often because they had no subjective problems but only occasional pain or skipping feelings. In 14 patients, the surgery of nonunion was performed by removing the intramedullary nail and plate stabilization. This group of patients was analyzed in more detail.

The mean age of 14 patients in the group was 64 years (31–84). The mean follow-up period after the revision surgery was 26 months (12–64). The original fractures in the group of operated nonunions were classified according to the AO classification. There were 5 type A1 fractures, two type A2 fractures, five type B1 fractures and two type C1 fractures. Six cases of nonunion were hypertrophic, and eight patients had arthophic nonunion (tab 1).

Tab. 1: Patient group

	Number (n)
Non-unions after intramedullary nailing	18
Sex
Man	5
Woman	13
Fracture type (AO/OTA)
Type A1	5
Type A2	2
Type B1	5
Type C1	2
Type of nonunion
Hypertrophic	6
Atrophic	8

The healing of the nonunion occurred in all operated patients in an average period of 4.1 months (2.9–7.7). The DASH score was 13 points at the end of the follow-up of the group of revision surgeries (1–32). Constant shoulder score was 77 points (53–98) on average. After the revision surgery, 3 patients had transient paresis of the radial nerve. In two cases, nerve function was fully restored. In one case, distal forearm and hand hyperaesthesia persist. The average elbow extension limitation is 18° (5–42). Average elbow flexion 123° (105–140) (tab. 2).

Tab. 2: Group results

Number of patients	14
DASH score	29 (1–72 points)
Constant score	77 (53–98 points)
Elbow extension restriction	18° (5°–40°)
Achieved elbow flexion	123° (105°–140°)
Partial paresis of n. radialis	3

DISCUSSION

The nonunions of the humeral bone occur after surgical treatment in 8  – 12 % [10, 29, 31]. They are more common in high-energy injuries with skeletal defects and open fractures [27]. Also, soft tissue interpositions or the presence of infection can lead to nonunion [26]. Only part of the patients undergo a revision surgery of the nonunion. Subjective complaints are not as severe as those of the long bones of the lower limb. Some elderly patients refuse the recommended surgery [20]. In relation to the bimodal distribution of humeral diaphysis fractures in the population, we also observe more frequent incidence nonunions in women after the sixth decade.When nailing the fracture of the humeral diaphysis, it is important not to leave distraction at the fracture site [9, 12, 28]. The nail is inserted into the reduced fracture. If the fracture reposition fails in a closed manner, direct reposition is performed. We introduce a nail from the miniincision and after the insertion of the reduction forceps into the reduced fracture. The use of a cerclage tape or cerclage wire may be beneficial. Metsemakers recommends the use of a tape in the event of dislocation of fragments over 2 cm [20]. Other authors reject cerclage and point to the possibility of neurovascular injury and the risk of infection [16]. We use tapes when the fracture cannot be reduced in a closed manner or from iniincision with reduction forceps.

In terms of the fracture type, long diagonal fractures of type 12 A1 and interfragment fractures of type 12 B, in particular spiral type 12 B1, appear to be the most risky for the development of nonunion. Castella describes fractures at the transition of the proximal metaphysis and the diaphysis of the humerus with long broken out fragment as predisposing to the nonunion. The metaphyseal part heals well, and an atrophic nonunion occurs between the distal part of the fragment and the diaphysis [7]. This is confirmed by our experience as well.

The gold standard in addressing the nonunions humeral diaphysis after nailing is currently the splint osteosynthesis with compression after previous nail extraction. The reported percentage of healing in the groups is from 85 to 100 % [7, 27, 32]. Nadkarni presents good results after leaving the nail in situ and increasing stability with the underposed plate [22]. Another option is to replace the nail with a stronger one with re-drilling of the medullar cavity. However, the results are not entirely convincing, because the benefits of axial loading as on the lower limb [6, 18] cannot be utilized. Another possibility is to solve the nonunion using a circular fixator with compression [25]. The authors achieve a high percentage of healing but with a higher percentage of complications. The success rate of healing of nonunion by non-invasive methods such as electrical stimulation or ultrasound is reported to be from 0 to 60 % [1, 23]. Hypertrophic nonunions generally arise with insufficient biomechanical stability. It can be restored with good bone quality by 4.5 mm autocompression plate. All the necessary factors for bone healing are present in the hypertrophic nonunion so ensuring the mechanical stabi­lity is sufficient for healing. We have also been successful in addressing hypertrophic nonunions with drilling or chipping technique by an oscillating saw to improve the vascularity of the nonunion [19]. Treatment of atrophic nonunions involves resection of the avital sclerotic bone up to the bleeding tissue even at the cost of bone shortening. Shortening of humerus up to 3 cm is well tolera­ted and has no functional response in terms of limiting of the elbow mobility. The intramedullary canal is trea­ted with drillings or excochleation with sharp spoon. This step is important because it brings osteoproductive cells to the site of the nonunion being addressed. The use of autogenous spongioplasty in atrophic nonunions is recommended [14, 15, 17].

Upon closer examination of our failed nails, we found that nail breakage was the most common with nails that have the same diameter along their entire length. Nails that are designed to be wider in the proximal part than in the distal part have never broken. On the contrary, with these unequally strong nails, we observed a breakage of “only” the locking screws due to instability. Another finding was that all broken nails are designed to have locking holes along their entire length. The break always occurred in one of these holes. These associations have not been discussed in the recent literature.

We did not record an infected nonunion in the group. Its solution is complex and depends not only on the local condition of the soft tissues but also on the general condition of the patient and the associated diseases. In extreme cases, the remediation intervention may be amputation. In most cases, a radical resection of the infected site is indicated along with a sufficiently radical debridement of soft tissues with subsequent secondary reconstruction on the skeleton and soft tissues [15, 32].

Case report 1

Case report 2

Case report 3

CONCLUSION

The solution of each nonunion should be approached individually with good preoperative planning. CT scan including three-dimensional reconstruction is appropriate. The vast majority of nonunions are due to a technical error. The most common error is the insufficient fracture reduction prior to nail insertion. Fractures of type 12A1 and 12B1 are more predisposed to origination of nonunion. In the case of nail-induced nonunion proven by X-ray, although patients do not have major subjective problems, we suggest a revision surgery to patients. If the nail or locking screws break, reoperations tend to be more complicated. Aseptic nonunions of the humeral diaphysis after intramedullary nailing can be reliably resolved by nail removal, site revision and plate osteosynthesis.

doc. MUDr. Radek Veselý, Ph.D.

r.vesely@unbr.cz