Increased wound complication with intramedullary screw fixation of clavicle fractures: Is it thermal necrosis?

Introduction

Clavicle fractures are common in adults, accounting for close to 3% of all fractures. Primary operative fixation has become increasingly popular for the treatment of displaced, comminuted mid-shaft clavicle fractures. ¹

In a systematic review, there was a lower overall complication rate in the operative group compared to the nonoperative group; however, this difference did not reach significance. ² Operative implants and techniques vary and include different intramedullary (IM) osteosynthesis and plate fixation. ^1,3,4 The commonest operative complications consisted of local hardware irritation or pin protrusion and wound problems or infection.

In the review by Zlowodzki et al., the infection rate for plating was 4.6% and the rate for IM pins was 6.6%. ⁵

Wijdicks et al. suggested that infection is defined as redness, swelling, purulent discharge, a positive wound culture, and/or when prescription of antibiotics was given, and irritation was caused by prominence or protrusion of the implant material. ⁶

There are only few articles ^{7 –13} reporting on these complications and even fewer ^14,15 explore their treatments and outcomes, especially with IM cannulated screws. ^16,17

Our study was designed to focus on the treatment and outcome of patients with clavicle fractures treated with open reduction and internal fixation, who had local wound complications. Our hypothesis was that there is no difference in this complication between different surgical fixations.

Materials and methods

Ninety-seven patients underwent operative treatment for a clavicle fracture between 2010 and 2013. There were 49 male and 48 female patients with the mean age of 38 years old at the time of injury; 40% of fractures were due to high energy (road traffic accident, fall from horse, and so on) with only three open fractures. All fractures were middle third clavicle fractures with at least 2 cm of shortening or displacement with no bony apposition or associated chest injuries. The fixation method was selected by the two senior operating trauma surgeons, according to the fracture configuration, the soft tissue envelope, and the patient’s condition; 17 patients were treated with open reduction and IM fixation (6.5-mm Arbeitsgemeinschaft für Osteosynthesefragen (AO) cannulated screw); 80 patients had open reduction and internal fixation with Acumed (Hillsboro, Oregon, USA) pre-contoured anatomic plate fixation, with or without lag screw.

All patients with plates were operated on using standard AO technique and an infraclavicular approach. Simple fractures were fixed with a lag screw and a superior neutralization plate to obtain absolute stability. In multifragmentary fractures, fixation was achieved with a bridging plate restoring length, alignment, and rotation.

During cannulated screw fixation, the fracture site was exposed through a small incision along Langer’s line. The IM canal of the medial fragment was drilled, and then, a guide wire was drilled retrograde through the lateral fragment and out the posterolateral aspect of the clavicle. The lateral fragment was drilled over the guide wire and after fracture reduction, the screw was advanced anterograde across the fracture site and into the medial fragment.

After surgery, all patients were given a sling for comfort (2–4 weeks) and started an early active mobilization program, although resistance with the limb was avoided for 6 weeks postoperatively. All patients had routine follow-up at 2, 6, and 12 weeks and at the final assessment.

Patients with local postoperative wound complications were assessed retrospectively at mean 58.3 months (31–60) after their initial treatment with visual analogue pain scores (VAS), new Oxford Shoulder Score (OSS), QuickDash (QD) Score, and anteroposterior and axillary view radiographs.

Results

Eight cases (8.2%) were diagnosed with postoperative wound complications (mean: 19.5 days (9–49)). The mean age of these eight patients (six males and two females) was 48.3 years (27–73) at the time of the injury with 4 left and 4 right side injuries; five of eight injuries were high energy, with only one open fracture (Table 1). These cases with wound complications were all operated on at mean 17.3 days (1–42) following their injuries. Patients with stable fixation were treated with the “dress and suppress method” (regular dressing, antibiotic treatment, and delayed metalwork removal if necessary). Patients with unstable fixation were treated with metalwork removal, debridement, and antibiotics.

Five patients (three IM and two plate fixation) had a wound breakdown, in only one was an organism grown initially (Coagulase negative Staphylococcus); three patients had wound erythema (two IM and one plate fixation). There were no predisposing factors for soft tissue complications found in these cases.

In seven patients with stable fixation, it was possible to “dress and suppress,” these were treated on average with 3 weeks (1–6) of oral antibiotics and had resolved symptoms at mean 8 weeks (3–15; Table 2); three of these seven patients had delayed metalwork removal due to irritation (two cannulated screws and one plate), at average 11 weeks after their initial surgery.

OPEN IN VIEWER

Table 1.

Patients with local wound complications after clavicle fixation.

Patient number	Gender/age	Side	Injury energy	Fixation type	Injury to surgery (days)	Wound complication/days after surgery	Microbiology
1	M/66	L	High (open)	Plate	1	Wound breakdown/14	Neg (initially) then Staphylococcus aureus
2	M/46	R	High	Plate	6	Erythema/19	Neg
3	F/73	R	Low	IM (screw)	24	Erythema/49	Neg
4	M/57	R	High	IM (screw)	13	Erythema/28	Neg
5	M/27	R	Low	IM (screw)	12	Wound breakdown/10	Coag Neg Staph
6	F/41	L	High	IM (4.5 mm screw)	27	Wound breakdown/12	Neg
7	M/42	L	High	Plate	42	Wound breakdown/9	Neg
8	M/34	L	Low	IM (screw)	13	Wound breakdown/15	Neg

Coag Neg Staph: Coagulase Negative Staphylococcus; F: Female; IM: intramedullary; L: Left; M: Male; Neg: Negative; R: Right.

OPEN IN VIEWER

Table 2.

Treatment and outcomes of wound complications.

Patient number	Treatment (weeks of antibiotics)	Symptoms treated at (weeks)	Further surgery or complications	Follow-up (months)	VAS	OSS	QD
1	2/52 oral	4/52	Chronic/recurrent infection with Staphylococcus aureus, late removal of metalwork at 28 months	31	1	43	6.8
2	3/52 oral	6/52	No	80	0	48	0
3	3/52 oral	13/52	Removal of metalwork at 14/52	58	2	45	6.8
4	1/52 oral	10/52	No	70	2	47	9.1
5	6/52 oral	3/52	Removal of metalwork at 6/52	57	0	48	0
6	2/52 i.v. and 6/52 oral	10/52	Unstable fixation, debridement at 2/52 and removal of metalwork at 3/52	53	3	46	4.5
7	1/52 oral	4/52	Removal of metalwork at 13/52	62	1	46	4.5
8	5/52 oral	15/52	No	56	1	46	4.5

VAS: visual analogue pain score; OSS: Oxford Shoulder Score; QD: QuickDash; i.v: intravenous.

One out of these seven patients, who had a plate fixation for an open fracture, failed this conservative method later and developed a chronic discharging wound with no evidence of osteomyelitis, requiring debridement and plate removal at 28 months. Intraoperatively, there was no evidence of macroscopic infection, histology confirmed granuloma, and microbiology revealed low-grade Staphylococcus Aureus infection, which was treated with oral antibiotics successfully.

One patient with a small and narrow clavicle had unstable fixation with a 4.5-mm cannulated screw (Figure 1). She had anterior wound breakdown, which was not suitable for the “dress and suppress method” and required 2 weeks intravenous and 6 weeks oral antibiotic treatment. She underwent debridement followed by early screw removal at 3 weeks and achieved bony union at 14 weeks.

OPEN IN VIEWER

Figure 1.

(a) Preoperative anteroposterior view and (b) postoperative anteroposterior view of the radiographs of the complicated clavicle fracture with unstable fixation (4.5-mm cannulated screw).

At the final follow-up, all wounds remained dry and healed. Clinical and radiographic union was achieved in all. The average VAS score was 1 (0–3). The mean OSS and QD scores were 46 (43–48) and 4.5 (0–9.1), respectively.

Although the timescale from injury to surgery, the patients’ demographics, and the final outcomes were similar in both treatment groups, and we used the same skin preparation, antibiotic prophylaxis, and sterile services, the wound complication rates differed significantly between IM and plate fixation, with 5 out of 17 (29.4%) in the IM group and 3 out of 80 (3.8%) in the plate group. In the IM group, all wound complications were anterior and not related to prominent metalwork posteriorly.

Discussion

Plate fixation is one the most widely used options by many surgeons, for the surgical treatment of displaced clavicle fractures. ^{1 –3,7,8} Several IM fixation options were also described, which can be applied by anterograde or retrograde methods and can be classified as IM canal filling (Rockwood pin, Hagie pin, cannulated screw, and so on) or IM canal splinting Elastic stable intramedullary nailing (ESIN nail) methods.

Some IM techniques have their own drawbacks, particularly related to the prominence of the metalwork once the swelling subsides, or implant migration. ^4,6,10 Rockwood clavicle pins appear to be as effective as plates in achieving union and maintaining length in both the acute and delayed setting ^{11 –13,18} IM cannulated screw is an effective load-sharing alternative, which has no protruding lateral end, only a smooth head. ^{16,17,19 –21}

Duncan et al. realized first that most papers document local wound complications, but there is a lack of outcomes and treatment protocols for the management of septic complications after clavicle fracture. ¹⁴ He identified five patients with positive microbiology results (three Propionibacterium and two Staphylococcus aureus). He recommended aggressive early debridement of all necrotic tissue and removal of all nonabsorbable sutures and implants, followed by intravenous antibiotics for 4–6 weeks, with additional oral antibiotics provided on an individual basis. Only two patients went on to have bony union and some patients had pain and limited function.

Liu et al. published an infection rate of 4.9% after reconstruction plate fixation and most of his cases healed with primary bony union, after debridement and early removal of the implants. ¹⁵ Böstman et al. also reported 7.8% infection rate following plate fixation and used early debridement and metalwork removal or replating in the majority of these cases for successful treatment. ⁸ Verborgt et al. documented 18% infection rate with plating but without any identified organism. ⁷ He had a less aggressive treatment approach as three patients were treated with early debridement and antibiotics and four required only antibiotic treatment to achieve bony union in all. In our study, we reported 3.8% wound complication rate with clavicle plating.

With 2.5-mm threaded pin, Grassi et al. published 20% infection rate and good outcome with conservative treatment. ⁹ Judd et al. reported 31% wound complications after modified Hagie pin fixation and most cases required debridement and premature metalwork removal. ¹⁰ Mudd et al. used Rockwood pin for fixation and documented 33.3% local soft tissue problems, with only one positive microbiological result (S. aureus), but most patients were treated with premature pin removal. ¹³ Marlow et al. published 17.4% wound problems and all were successfully treated with delayed Rockwood pin removal. ¹¹

These recently published increasing local wound complications were also described in our results, with 5 out of 17 (29.4%) in the IM cannulated screw group, although the timescale from injury to surgery, the patients’ demographics, the final outcomes, and the intraoperative and postoperative details were similar in both fixation groups. All our IM screws were inserted laterally using a similar surgical technique described by several other studies. Abo El Nor ¹⁶ reported 5%, Richardson et al. ¹⁷ published 7%, and Krishnan et al. ¹⁹ documented 20% local soft tissue problems following cannulated screw fixation, without any microbiological evidence of infection.

Khalil reported his results of 37 patients with medially implanted 6.5-mm partially threaded cancellous screws. ²² Superficial infection was observed only in two diabetic patients and was controlled with antibiotics and dressing. The medial starting point appeared to decrease the incidence of implant prominence and skin breakdown.

Sun et al. published no complication with IM screws, but 4% poor wound healing following plate fixation, which were treated with the intermittent removal of stitches and regular dressing changes. ²⁰

In recent reviews by Wijdicks et al., they reported wound and deep infection rates below 10% for plate fixation, and the majority of these wound infections were treated successfully with oral antibiotics. ²³ He reported major complication rates for IM fixation to be no higher than 7%. Most wound complications in the IM group resolved with antibiotics and/or shortening/removal of the IM device. ²⁴

It is also important to note that the clavicles of women with slight figures may be too thin to use 6.5-mm cannulated screws; in such cases, 6.00 mm or even thinner ones can be tried. ²⁰ If less than 1 cm of the medullary cavity can be fully fixed at the proximal end of the clavicle, IM cannulated screw fixation should be abandoned. Otherwise, it can lead to unstable fixation, with increased risk of soft tissue complications and further surgery, as in one case in our current series.

One of our group previously published the overall outcome after clavicle fixation of the same group of patients at mean 31 months follow-up. ²⁵ The mean age of the patients at the time of injury was 38 years, which is 10 years younger than in this current predominantly male group of patients with local wound problems. The rate of wound complication was comparable in the early and delayed fixation groups in this initial publication, which suggests that maybe increased age and male gender can be risk factors for this problem. The initial mean OSS and QD score were 44 (compare to recent 46) and 9 (compare to recent 4.5), respectively, which indicates further functional improvement, even a few years after a complicated clavicle surgery.

The high number of soft tissue complications that occurred in several series following IM fixation deserves further careful consideration. The subcutaneous position of the IM device contributes to the overlying skin necrosis posteriorly, which is clearly a concern for bacterial seeding of the implant, although deep infections or microbiologically confirmed infections are rarely reported in literature. ^13,15 We also found only two patients with positive microbiology results, which may suggest a primary inflammatory process rather than an infective one. The senior surgeons became suspicious about the apparent increase in wound complications and during the last IM fixation undertaken on a young male, noted blackening and heating of the guide wire used for retrograde drilling. The serial drilling of thick cortex (especially the posterolateral) during IM canal preparation and the limited soft tissue cover of the subcutaneous clavicle may lead to inadequate heat dissipation. Our theory is that the difference in our complication rates compared to the recently published increased local wound complications mostly without microbiological evidence may be due to the IM technique of drilling with a guide wire and due to aseptic thermal bone necrosis, rather than true infection. During IM cannulated screw fixation of clavicle fractures, we would advise drilling retrograde with a sharp drill, with continuous irrigation, and being cautious of the dense compact bone at the lateral aspect of the clavicle with poor facility for heat dissipation.

Our study has limitations. This was a small retrospective study, and due to the surgeons’ preference of treatment method, we also encountered a difference in treatment policy for different types of fractures. Patients were not randomized to determine the type of fixation, so simple fractures and polytrauma patients had a greater possibility of being treated with IM fixation than complex fractures. However, the main goal of this study was to identify all cases and document the outcome of patients with clavicle fractures treated with internal fixation, who had local wound complications.

Conclusions

We can conclude that a good outcome of a dry healed wound, united clavicle, and good function can be achieved in most patients with the “dress and suppress” method, without the need of early aggressive debridement and implant removal (provided the initial fixation is sound). However, further studies with a larger sample size are required in future, to confirm the outcome achieved and our possible theory of aseptic bone necrosis.