Hip arthroplasty in failed intertrochanteric fractures in elderly.

BACKGROUND: Failed intertrochanteric fractures in elderly patients are surgical challenge with limited options. Hip arthroplasty is a good salvage procedure even though it involves technical issues such as implant removal, bone loss, poor bone quality, trochanteric nonunion and difficulty of surgical exposure.
MATERIALS AND METHODS: 30 patients of failed intertrochanteric fractures where hip arthroplasty was done between May 2008 and December 2011 were included in study. 13 were males and 17 were females with average age of 67.3 years. There were 2 cemented bipolar arthroplasties, 19 uncemented bipolar, 4 cemented total hip arthroplasty and 5 uncemented total hip arthroplasties. 16 patients had a trochanteric nonunion, which was treated by tension band principles. Total hip was considered where there was acetabular damage due to the penetration of implant.
RESULTS: The average followup was 20 months (range 6-48 months). Patients were followed up from 6 to 48 months with average followup of 20 months. None of the patients were lost to followup. There was no dislocation. All patients were ambulatory at the final followup.
CONCLUSION: A predictable functional outcome can be achieved by hip arthroplasty in elderly patients with failed intertrochanteric fractures. Though technically demanding, properly performed hip arthroplasty can be a good salvage option for this patient group.

INTRODUCTION

Intertrochanteric fractures in elderly population are usually treated by closed reduction and internal fixation. There have been many advances in internal fixation techniques to deal with poor quality of bone and severely comminuted intertrochanteric fractures so as to allow early ambulation.[1] Most of these patients have good results.[2] Failure rates with an internal fixation range between 3% and 12%; device penetration between 2% and 12%, nonunion 2-5% and malunion with varus deformity in 5-11%. Certain unstable fracture patterns have been reported to have failure rates[3] as high as 56%.[4] Revision of internal fixation for nonunited fractures have been reported to have good results in younger patients with good bone stock.[5678] In older patients with poor bone quality and degenerative acetabular changes, hip arthroplasty has been advocated as a salvage procedure.[39] There have been a number of technical issues to convert these fractures into hip arthroplasty such as extraction of implants, bone deformity, bone loss, poor bone quality and associated trochanteric nonunion.[101112131415]

The purpose of this retrospective study was to evaluate clinically and radiologically, the early results of hip arthroplasty in this elderly group of patients with failed intertrochanteric fractures. This study also describes the complications and technical difficulties while operating this subset of patients.

MATERIALS AND METHODS

After obtaining approval from institutional review board we reviewed retrospective data of all 30 patients who had under gone hip arthroplasty for failed intertrochanteric fracture between May 2008 and December 2011. Of these 30 fractures, there were 26 failed intertrochanteric fractures and 4 were failed intertrochanteric fractures with subtrochanteric extension. There were 13 males and 17 females with an average age of 67.3 years (range: 51-81 years). The average body mass index of these patients was 24 [range 18 to 28].

Sliding hip screw (n = 13), proximal femoral nail (n = 6), Enders nail (n = 3), Smith-Peterson nail plate (n = 2), angle blade plate (n = 2), 95° sliding screw (n = 2), Jewett nail plate (n = 1) and locking Cobra plate (n = 1) were used to treat these fractures primarily. The average time between the fracture and arthroplasty was 44.6 months (range 4-324 months). 11 patients underwent multiple surgical procedures before the hip arthroplasty; the average being 1.5 procedures (range 1-4). All fractures had nonunion, which was assessed radiologically, clinically and using computerized tomography in doubtful cases.

All patients had undergone preoperative clinical examination and medical evaluation for their co-morbidities. The operative notes, hematological investigations and their preoperative radiographs were collected from the database. For occult infections, a complete blood count, erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) was taken into consideration. The comorbidities noted were hypertension (n = 23), diabetics (n = 15), ischemic heart disease (n = 5), hypothyroidism (n = 2) and cerebrovascular accident (n = 1).

Operative procedure

In patients with proximal femoral nail and Enders nails, full length radiographs of the femur were obtained to know the site of entry of Ender's nail or the position of the locking screw. All possible inventories for implant removal were kept available including high speed burrs and hollow mill. Universal extractor was used for removing proximal femoral nails. All patients were operated by a single surgeon in lateral decubitus position using the posterior approach. Previous scars were not taken into account in planning new incisions. In all cases, tissue culture was sent from around the implant. Femoral head was dislocated before implant removal. In patients who had cut out of the implant, the implant was removed before dislocation. Two patients needed a hollow mill for removal of broken shaft screws. The complete head and neck was removed en mass keeping the posterior capsular flap for closure. Out of these 30 patients, 16 patients were having nonunion of greater trochanter. Surgical difficulty was greater in cases where there was a greater trochanter nonunion, due to fibrous tissues and malpositioned greater trochanter. Acetabular labrum was never excised except in cases where total hip arthroplasty was being performed. Patients with acetabular cartilage damage due to the implant penetration underwent total hip arthroplasty.

A total of 9 patients underwent total hip replacement of which there were 4 cemented and 5 uncemented hip replacements. A 28 mm metal head was used in all cases. In 21 patients, bipolar hip arthroplasty was performed. 2 of them underwent cemented femoral fixation with a standard stem. 19 of them underwent uncemented femoral fixation of which 2 were with proximally coated stems and 17 with fully coated stems. Four patients in whom fully coated stems were used had calcar replacing stems. Stem selection for patients depended upon available bone quality and stability of trial on table.

Most of these patients had blocked canal due to nonunion and fibrosis. To access the canal, the initial drilling was followed by short intramedullary cannulated reamers over the guide wire to ream the canal. There was no metaphyseal cancellous bone to prepare the canal with routine broaches; hence, we had to use larger reamers and high speed burr in the proximal part of the canal to accommodate the prosthesis. In one patient, who had subtrochanteric extension with 4 prior surgeries had proximal fragment instability. Hence, lateral plate was used to stabilize the proximal fragment and bone grafting was carried out on the posteromedial side of the femur.

Of 30 patients, 16 patients had greater trochanteric nonunion (53.3%). In all these patients, the trochanteric-nonunion was mobilized, fibrous tissue was debrided and high-speed burr was used to decorticate the bone. The trochanter was brought to the best possible anatomical position and was wired with the tension band principle [Figure 1]. The horizontal wire of Charnley was used only in cases where there was adequate proximal bone available. In all cases, 2 k-wires from the trochanter to the proximal shaft and figure of 8 tension band on these wires were done [Figure 1]. This added stability in mediolateral plane.

Figure 1

(a,b) X-ray anteroposterior and lateral views of hip with thigh in an 80 year old female with 6 months old failed intertrochanteric fracture with enders nail and cannulated screws in situ. (c) Preoperative computed tomography scan shows nonunion of intertrochanteric fracture and trochanteric nonunion. (d) Immediate postoperative radiograph anteroposterior view shows fully coated stem with distal fixation and trochanteric wiring. Postoperative 2 years anteroposterior (e) and lateral (f) radiograph showing union of trochanter with no subsidence of femoral stem

Image intensifier was used when needed. Patients with cephalomedullary nails had great difficulty in broaching the canal due to the trochanteric entry and sclerosis around the nail. High speed burr was used to clear this sclerosis.

Limb length was only adjusted by soft tissue tension as the local anatomy was distorted. Pyriformis and posterior capsule was closed with drill holes through the greater trochanter. Drains were kept in situ for 48 h. All patients were given chemoprophylaxis and compression stockings to prevent deep vein thrombosis (DVT). Patients were ambulated on the 3rd postoperative day with the help of a walker. Weight bearing according to the patient's tolerance was allowed. Patients were discharged on the 6th or 7th postoperative day.

All patients were followedup clinically, radiologically at 2 weeks, 6 weeks, 3 months and annually thereafter. Of the thirty patients 26 had minimum followup of 1 year or more while 4 had a followup of less than 1 year. Independent observers used Harris hip score, limb length measurement, visual analog score and radiological documentation. We used the Harris hip score[16] to follow the patients clinically. For cemented arthroplasty, we used radiological criteria to assess lysis according to Gruen's zones.[17] For uncemented arthroplasty, the criteria proposed by Engh et al. were used.[18] Heterotopic ossification was graded according to Brooker's grading.[19] Trochanteric union was assessed. Patient's satisfaction was assessed using the visual analog scale with the worst score being zero and the best score being ten. For bipolar hip arthroplasty, acetabular wear was not measured as the followup was too short. None of the patients were lost to followup. Of 30 patient's 26 were assessed clinically and radiologically by the operating surgeon and an independent observer. Four patients had telephonic interview for assessment for satisfaction and their latest X-rays were obtained through mail.

RESULTS

30 patients were followed up for an average of 20 months (range 6-48 months). All patients had normal ESR and CRP preoperatively and none of the intraoperative cultures grew organisms. We had no mortality until latest followup. There were no dislocations. Two patients (6.7%) had DVT proven by Doppler and were treated pharmacologically. An average of 2.5 units of blood per patient was given. One patient had wound dehiscence, which was debrided and sutured under local anesthesia. The wound had healed at the latest followup.

Harris hip score increased from a preoperative mean of 27.9 to a postoperative mean of 70.6 (range 53-88) postoperatively. All patients demonstrated significant pain relief and were ambulatory. A total of 12 patients (40%) are walking with the help of a cane, three (10%) with the help of a walker and 15 patients (50%) are walking without any aids. None of the patients in our study had trochanteric pain. One patient (3.35%) had anterior thigh pain, which settled down with supportive measures after 3 months. Of 30 patients, sixteen (53.3%) had trochanteric nonunion preoperatively. Of the 16 patients, four patients had nonunion (fibrous) post arthroplasty. None of these patients had migration of the greater trochanter of more than 15 mm postoperatively. Three of four patients with trochanteric nonunion needed a walker for ambulation while one needed a stick. These patients had lower Harris Hip scores (mean 60.7 range 53-66 as compared to those who achieved union. The remaining 12 patients showed bony union. Of these four patients were able to walk without support while eight needed a cane for ambulation. Two patients (6.7%) had breakage of wires. None of the patients had heterotopic ossification. In this short followup of an average of 20 months, no implant was unstable or has any signs of loosening to consider revision. Five patients had shortening of 1 cm postoperatively. Radiologically, there was no significant subsidence. At the final followup, the patient satisfaction scale by visual analog had an average of 9.1 (range 8-10).

DISCUSSION

Patients with failed internal fixation of intertrochanteric fractures present with significant functional disability and pain. In these elderly patients with disuse osteopenia treatment options allowing early ambulation are limited. Hip arthroplasty can provide a good salvage option for early ambulation and restoration of normal life in this group of patients.

The authors acknowledge the drawbacks of a retrospective study. The number of patients was small with limited followup and had failed intertrochanteric fracture fracture treatment due to a variety of implants. It was not possible to evaluate the fracture patterns and causes of failure. A variety of treatment options were used to manage these patients ranging from modular bipolar (cemented and cementless) to total hip replacement (cemented and cementless), which makes comparison of results in these heterogenous group of patients difficult. However the decision to use a particular implant is based on the individual patient age, bone stock and type of implant failure and it is difficult to use a single modality to treat these patients. 21 patients received a modular bipolar prosthesis and 4 patients with acetabular erosion received total hip replacement. While selecting hip arthroplasty as a treatment option for these patients it must be kept in mind that this procedure is technically demanding for this particular class of patients. Poor bone quality, loss of bone stock, presence of holes after removal of hardware combined with distorted bony landmark and severe fibrosis can increase the risk of mechanical complication like fractures and cortical perforation. When trochanteric nonunion is present, there is usually severe fibrosis at the site leading to difficulty in exposure. These patients need exposure around the head and neck with 360° visualization of acetabular before extracting the head and neck fragment. It is mandatory to keep the capsular flaps and labrum whenever bipolar arthroplasty is being performed. Patients with cephalomedullary nails in situ have neocortex formation, which makes the canal preparation challenging. In those patients without intramedullary device, the canal is blocked by fibrous tissue due to nonunion, which needs careful opening. There is no proximal metaphyseal bone for stability of the implant. In such cases, implants, with distal diaphyseal fixation and must be considered fully coated stems. There is a mismatch between the proximal and distal configuration; hence, most cases need distal fixation. Modular implants may have an advantage in such a situation. Re-attachment of the trochanter in patients with nonunion is a crucial step to gain stability and reduce the rate of dislocation. The original two wire technique of Sir John Charnley, which includes two vertical wires and one horizontal wire whenever possible was used. Two K-wires with a figure of eight tension band wiring without horizontal wire in patients with no proximal medial bone was used. Twelve of sixteen patients had greater trochanteric union. Four patients had nonunion, but did not have severe migration of the greater trochanter. We feel that trochanteric union may not be achieved in all the cases; however, an intact soft tissue sleeve including the gluteus medius and vastus lateralis will prevent trochanteric escape and henceforth prevents instability. 15 patients had abductor lurch and needed walking aids for walking. The additional stability can be obtained by enhanced posterior capsular repair.

The choice of implant-whether cemented or uncemented, total hip replacement or bipolar arthroplasty will depend upon the age of the patient, activity level, co-morbidities, quality of bone stock and acetabular damage. Patients with acetabular damage due to implant penetration underwent total hip replacement [Figure 2a–e]. We feel the low reduction rate of dislocation in our series, compared with other published series were due to the good union rate of greater trochanter. Cemented stems were used in patients with good proximal bone where adequate cementing was possible with standard stems. All other patients except those mentioned above underwent bipolar hip arthroplasty with fully coated distal fixation stems [Figure 3]. We used calcar replacement, fully coated stems in four patients who had loss of proximal bone including the lesser trochanter.

Figure 2

(a) Anteroposterior and (b) lateral radiograph of 72 years female with failed proximal femoral nail done for intertrochanteric fracture femur. Proximal two lag screws were removed due to penetration and cut out. (c) Immediate postoperative cemented total hip replacement with acetabular graft done for the defect in posterior-superior acetabular region due to penetration of the screws. (d) Two years postoperative anteroposterior and lateral (e) followup radiographs showing good graft incoporation

Figure 3

Preoperative anteroposterior radiograph showing (a) Failed intertrochanteric fracture with broken 95° angled blade plate (b) Immediate postoperative anteroposterior radiograph shows fully coated stem used to bypass stress riser screw holes. (c) Postoperative anteroposterior (c) and lateral (d) radiographs showing no subsidence of stem

Haidukewych and Berry[9] studied the results of hip arthroplasty in 60 patients of failed internal fixation. Of the 60, there were 32 total hip arthroplasty and 28 hemiarthroplasty. At a mean followup of 65 months, 39 (89%) of the 44 surviving patients had no or mild pain. 91% were able to walk; 59% were ambulating with one arm support or less.

Zhang et al.[20] have reported a high incidence of complications after 14 cemented, 3 uncemented and 2 hybrid arthroplasty for failed internal fixation of intertrochanteric fractures. Their overall complication rate was 47% with 31% incidence of intraoperative fracture and postoperative dislocation in 16%. The mean Harris hip score at 2 years was 79.8.

Srivastav et al.[21] studied total hip arthroplasty following failed internal fixation of proximal hip fractures. Their study included 10 intertrochanteric and 2 sub-trochanteric fractures. They had an average followup of 4 years over, which the mean Harris hip score increased from 32 to 79 postoperatively. All patients were ambulatory with one patient (5%) using a walker, four patients (20%) using a cane while the remaining walk without the help of any aids. There were 3 dislocations (14.3%), 3 intraoperative fractures (14.3%), 1 superficial and 1 deep infection in their study.

Thakur et al.[22] in their study of hip arthroplasty (12 total hip and 5 bipolar hip arthroplasty) for failed intertrochanteric fractures using a modular distally fixing stem had an increase in mean Harris hip score from 35.9 to 83 postoperatively. Two patients (13.3%) were walking with no aid, 10 (66.7%) were using a cane and the remaining 3 (20%) were ambulating using a walker. There were no dislocations, one patient (6.67%) had DVT and seven patients had heterotopic ossification (5 Brooker's grade 1 and 2 grade 2). In their study, a constrained liner was used in two patients as the stability of the hip intraoperatively was questionable.

To conclude, hip arthroplasty for failed intertrochanteric fractures requires meticulous preoperative planning and surgical technique. Posterior hip exposure makes the procedure easy for removal of implant additionally it allow safe accessibility to the nonunion of greater trochanter and intertrochanter which lies in mid of scarred tissue due to the previous surgery. Inventory like universal nail extractor, DHS removal set along with hollow mill and high speed burr are mandatory. Anatomic re-attachment of the greater trochanter is one of most important determinants of stability in these hips along with posterior capsular closure. The fact that we have had no dislocations in our study can be attributed to this technique. Implant selection depends on age of patient, level of activity, co-morbidity, condition of the acetabulum and available proximal bone. Distal fixation was needed in those cases where there was compromised proximal bone. Bipolar arthroplasty is an excellent option in elderly patients with co-morbidities without acetabular damage. All patients were allowed to bear weight postoperatively as tolerated and all were ambulatory at final followup. Thus, hip arthroplasty is a good alternative option for these failed intertrochanteric fractures.