Infected periprosthetic fractures around total hip arthroplasties are increasingly common and extremely challenging problem. The purpose of the study was to review the experience of two tertiary referral units managing infected periprosthetic femoral fractures using interlocking long-stem femoral prostheses either as temporary functional spacers or as definitive implants.

We have identified 69 patients with Vancouver B1 periprosthetic fractures around stable femoral implants. Open reduction internal fixation is the recommended treatment; however recent studies have revealed high rates of nonunion. We have reviewed the fixation techniques utilized to treat these patients, and identified outcomes in relation to rates of union, further surgery and mortality.

Patients were identified from a prospective database of all trauma admissions at Nottingham University Hospitals from 1999 to 2010. Hospital notes were independently reviewed and data retrieved.

69 patients were identified. Mean age 77 years and 63% were female. 51 (74%) occurred around total hip replacements and 18 (26%) around hip hemiarthroplasty after a previous hip fracture.

Periprosthetic fracture occurred around an Exeter stem (n=18), Charnley (n=10), Austin Moore (n=15), other (n=6). 20 patients had undergone previous revision surgery. The mean time from index surgery to fracture was 58 months (median 24) around primary stems, and 48 months (median 22) around revision stems.

6 patients (9%) were treated non-operatively. Five of these had undisplaced fractures (all healed but one required revision due to loosening) and one was too unwell.

63 patients (91%) were treated by open reduction internal fixation. Of these, single plate fixation was performed in 40 cases (64%). In the vast majority of cases, lag screw fixation of the fracture with a long (>12 hole) pre-contoured 4.5mm locking plate was utilised with cables. Both locking and cortical screws were used to achieve stable fixation (Figure 1). A double plate technique was used in 16 cases (25%), where plates were placed perpendicularly to each other (laterally and anteriorly). Strut grafts were used in 13 cases (21%). 7 patients (11%) were treated with cables alone.

23 patients have subsequently died (33%). Two have been lost to follow up and three are awaiting union. There is a mean follow-up of 35 months.

Deep infection occurred in 4 cases (6%). Non-union occurred in four cases (6%). Two of these were infected and one was treated with cables alone. Malunion occurred in one case treated with cables. One patient had a dislocation and two superficial infections occurred.

Further surgery took place in 8 patients (12%). Three of the infected cases were revised and one underwent wound washout. The two other non-unions were revised. Three further revisions were performed: one for malunion, one for aseptic loosening (treated non-operatively) and another for a second periprosthetic fracture. Superficial wound washout was performed in one case.

Mortality was 10% at 3 months, 22% at 1 year and 47% at 3 years.

We have identified that union can be achieved in the majority of cases after periprosthetic fracture fixation. Cable fixation was associated with a high complication rate (7 cases: two requiring revision surgery: one nonunion, one malunion).

We recommend that Vancouver B1 periprosthetic fractures are treated with meticulous technique to achieve anatomical reduction and fracture compression using lag screw technique and plating. Further mechanical support can be provided through the use of a second plate, cables and/or strut grafts.

The clinical results of the cemented Exeter stem in primary hip surgery have been excellent. The Exeter ‘philosophy’ has also been extended into the treatment of displaced intracapsular hip fractures with ‘cemented bipolars’ and the Exeter Trauma Stem (Howmedica).

We have identified an increase in the number of periprosthetic fractures that we see around the Exeter stem. We have also identified a particular group of patients with comminuted fractures around ‘well fixed’ Exeter stems after primary hip surgery that present a particular difficult clinical problem.

Prior to fracture, the stems are not loose, the cement mantle remains sound and bone quality surrounding the reconstruction is good, i.e. classifying it as a Vancouver B1. However the comminuted nature of the fracture makes reduction and fixation with traditional methods difficult. Therefore in these particular circumstances it is often better to manage these as B2 or even B3 fractures, with distal bypass and uncemented reconstruction.

Over an eleven-year period since 1999, 185 patients have been admitted to Nottingham University Hospitals with a periprosthetic femoral fracture around a hip replacement. These patients were identified from a prospective database of all trauma patients admitted to the institution. Of these patients we have identified a cohort of 21 patients (11%) with a periprosthetic fracture around an Exeter polished stem. Hospital notes were independently reviewed and data retrieved. Outcome data was collected with end points of fracture union, re-revision surgery and death. Data was also collected on immediate and long term post-operative complications.

The mean age was 76 years at time of fracture, and 52% were male. The mean duration between primary index surgery and fracture was 18 months (median 11 months).

15 patients were classified as Vancouver B1, and six as B2 fractures. Of the B1 fractures, 14 underwent fixation and one was treated non-operatively. Of the B2 fractures, four were revised, one was revised and fixed using a plate, and one was fixed using a double-plating technique. Prior to fracture, none of the implants were deemed loose although one patient was under review of a stress fracture which subsequently displaced.

One patient died prior to fracture union. All the other patients subsequently went onto unite at a mean of 4 months. There were no deep infections, non- or malunions. No patient underwent further surgery. Dislocation occurred in one patient and a superficial wound infection occurred in one patient which responded to antibiotic treatment. Three other patients have subsequently died at seven, twelve and fifty-three months post fracture due to unrelated causes.

In our series of patients, in addition to the more standard fracture patterns, we have identified a very much more comminuted fracture. Indeed, we have described the appearance as if the tapered stem behaves like an axe, splitting the proximal femur as a consequence of a direct axial load. As a consequence of the injury, the cement mantle itself is severely disrupted. There is significant comminution and soft tissue stripping, calling into question the viability of the residual fragments. Treatment of this type of fracture using a combination of plates, screws and cables is unlikely to provide a sufficiently sound reconstruction. In our experience we believe these fractures around previously ‘well fixed’ Exeter stems should be treated as B2/B3 injuries.