Conjoint Removal of Hip Screw–Femur Head during Hip Replacement after Previous Dynamic Hip Screw Fixation

Abstract

The purpose of this study was to demonstrate the reduced chances of iatrogenic femoral neck fracture while removing the Richard's screw using the dynamic hip screw–femur head conjoint removal technique. This retrospective cohort study analyzed 16 hips operated on with total hip arthroplasty from March 2010 to February 2015. All cases were previously treated with dynamic hip screws (DHS) for proximal femur fractures. The age of the patients ranged from 20 to 75 years. We used uncemented sockets in 15 patients and cemented sockets in 1 patient. We used conical fluted straight stems in 9 cases, ML (Mediolateral) tapered stems in 5 patients and CLS (Cementless Spotorno) stems in 2 patients. The head of the femur was removed together with the attached Richard's screw after taking a neck cut during hip replacement after previous dynamic hip screw fixation. At 2‐year follow‐up, there was a statistically significant improvement in the Harris hip score: from a mean preoperative score of 35 ± 7.975 to a mean postoperative score of 89.38 ± 4.870 (P < 0.001). Stem sinking and Type A_L (Vancouver classification for periprosthetic fracture) periprosthetic fracture in 1 patient with a tapered stem was noted. Good acetabular inclination was achieved in all cases. At 2‐year follow‐up, all patients were able to carry out their daily activities. This is a novel technique with the advantage of avoiding iatrogenic femoral neck fracture in an osteoporotic bone.

Introduction

Today, hip fracture surgery represents a large portion of the orthopedic surgeon's activity, and it tends to be associated with major clinical and social costs1. Over 90% of hip fracture patients are more than 65 years old and have preexisting medical comorbidities1. Both factors have an important influence on prognosis and treatment1. Even with optimal care, elderly trauma patients suffer higher morbidity and mortality rates when compared with the general population, and often require expensive hospital aftercare1. Due to these factors, surgical treatment of hip fractures in these patients are associated with unexpected clinical challenges. Hence, careful strategies are required for optimizing patient care1. Various methods of fixation are available for intertrochanteric femur fractures based on their subtype2. The dynamic hip screw (DHS) technique has proved to be a highly successful method for fixation of stable inter‐trochanteric fractures. The most common mechanical failure of fixation in using the sliding hip screw system is the cutout of the implant from the femoral head. Treating intertrochanteric fractures with a DHS may be associated with various complications, such as loss of reduction, nonunion, and malunion with varus deformity of the femoral neck, marked shortening of the affected limb, or screw cutout. Several authors have described modes of failure, types of complications of DHS fixation, and salvage procedures for failed fixation3, 4, 5. However, there are no articles on conversion of well‐fixed DHS with secondary arthritis of the hip to total hip arthroplasty (THA).

Considering the numerous complications associated with inter‐trochanteric femur fracture fixation with DHS, there are patients who end up with arthritis and eventually require a THA. Removal of Richard's screw in these cases produces shearing forces at the femoral neck and increases the chances of iatrogenic fracture. In addition, most of these patients are elderly patients with osteoporosis. Hence, the purpose of the present study was to bring to light a new technique of DHS removal, which has not been described in the published literature before, to reduce the chances of iatrogenic femoral neck fracture, along with the complications and challenges faced while performing this novel technique.

Materials and Methods

Inclusion and Exclusion Criteria

Inclusion criteria: (i) patients with well‐fixed DHS in‐situ who presented with hip pain; and (ii) patients with hip arthritis found by radiological examination.

Exclusion criteria: patients with fracture non‐union and with screw backout.

Patients’ Information

This study was approved by the institutional review board. We analyzed 16 hips treated with total hip arthroplasty from March 2010 to February 2015. The fractures had healed and patients were ambulating before they developed arthritis. The average duration between the first surgery and THA was 7.44 years (range, 2–30 years). Patients were aged between 20 and 75 years, with a mean age of 48.81 years. Ten patients (6 males and 4 females) had developed arthritis due to avascular necrosis and collapse of the femoral head. The remaining 3 male patients had arthritis and radiographs suggested cam type of primary osteoarthritis due to reduced head to neck ratio on account of irregular osseous prominence at the head–neck junction. The other 3 female patients had hip screws partially protruding from the anterior part of the femoral head that caused friction with the acetabulum and created a defect in the acetabulum. We used uncemented sockets in 15 patients and cemented sockets in 1 patient. We used conical fluted straight stems in 9 cases, SROM (Sivash Range Of Motion, DePuy Synthes, Warsaw, IN, USA) in 1 case, ML (Mediolateral, Zimmer, Warsaw IN, USA) tapered stems in 5 patients and CLS (Cementless Spotorno, Zimmer Biomet, Warsaw, IN, USA) stem in 1 patient (Table 1).

Table 1.

Individual patients’ information

Age (years)	Gender	Duration between the first surgery and THA (years)	Follow‐up period (years)	Implant	Acetabular inclination (°)	Harris hip score (preop)	Harris hip score (postop)
59	Female	1.5	6	CLS	46	29	82
48	Male	18	4	WAGNER	44	28	83
65	Male	1.5	5	SROM	40	34	91
20	Male	3	6	WAGNER	42	40	95
42	Male	7	4	ML	40	46	96
75	Female	2	8	WAGNER	50	27	90
38	Female	4	5	WAGNER	48	22	81
51	Female	4	6	ML	50	28	87
55	Male	30	2	WAGNER	44	33	90
61	Male	10	5	WAGNER	46	36	91
50	Female	3	3	ML	41	37	89
64	Male	6	5	ML	48	44	92
35	Male	3	4	WAGNER	46	44	95
21	Female	7	6	WAGNER	46	47	91
33	Male	11	6	WAGNER	46	40	94
64	Female	8	4	ML	48	25	83

CLS, Cementless Spotorno; ML, Mediolateral; SROM, Sivash Range of Motion; THA, total hip arthroplasty.

Surgical Technique

Anesthesia

All patients were given a combined spinal anesthesia (2.8 mL of 0.5% bupivacaine + 90 mcg buprenorphine) + epidural anesthesia (3 mL of 2% lignocaine + adrenaline 1:50 000) to help with pain alleviation and early ambulation after the surgery.

Position

All patients were put in lateral position with the operative limb facing upwards. Adequate support was applied to the sacral region and the lower abdomen to fix the position of the patient to determine the correct femoral anteversion and acetabular inclination while reaming the acetabulum and inserting the trial components.

Approach and Exposure

A posterior approach was used in all patients. The capsule with short external rotators was elevated from the posterior aspect of the trochanter. The vastus lateralis was separated and retracted to the anterior aspect to expose the Richard plate. Cortical screws were removed from the barrel plate.

Conjoint Removal Technique

A DHS is not a reverse cutting screw. As cancellous bone grows over and becomes compressed over it over the years, turning the screw in a reverse direction causes enormous shearing forces around the neck of the femur. There is a risk of spiral fracture of the neck. The technique involved dislocating the head of the femur and making the neck resection posteriorly, and using osteotome around the screw in sequence. We did not try to cut the screw but we cut the neck circumferentially around the screw. Then the head of the femur was removed posteriorly during regular total hip replacement with the dynamic screw attached to it (Fig. 1). The barrel plate came out from the lateral aspect of the trochanter and the proximal femur after the Richard's screw was removed. This technique has not been reported in previous studies, and we named it the DHS–femur head conjoint removal technique as we removed the DHS before removing the barrel plate. This technique helps in removing the head without damaging the residual short neck (Fig. 2).

Figure 1.

The patient was a 51‐year‐old woman who had developed right hip pain for 2 years. The patient had difficulty squatting and sitting cross‐legged. (A) Plain radiograph of the right hip anteroposterior (AP) view showing right hip arthritis with dynamic hip screws (DHS) in‐situ. (B) Intraoperative picture shows that the hip has been dislocated. (C) Head removal was done after taking the neck cut with the screw in‐situ. (D) Postoperative AP plain radiograph of pelvis with both the hip joints taken 1 day after the surgery showing implant in‐situ with adequate calcar with no radiographic evidence of iatrogenic fracture.

Figure 2.

Surgical diagram of conjoint removal technique. (A) Patient in lateral position with adequate padding of all bony prominences to avoid compression neuropathy. (B) Posterior approach taken to left hip where the short external rotators are divided and hip joint is visualized. (C) Capsulotomy done and hip joint dislocated. (D) Femoral neck cut made and Richard's screw exposed in‐situ. (E) Richard's screw removed along with the attached femoral head after completing the neck cut.

Intraoperative Observation

Two cases had greater trochanter fragments (approximately 1 cm × 0.5 cm and 1 cm × 1 cm respectively), which were separate above the barrel plate. The barrel plate had caused an iatrogenic fracture of the greater trochanter. The fragments had united by fibrous union. After removing the barrel plate, the trochanter was held separately and reattached to the stem after implantation using the tension band principle through holes in the stem.

We had to use a carbide drill bit to break the screw head that was bald and to remove the plate in three cases, leaving the screw in‐situ. Furthermore, we used a hollow mill to extract the screws. During removal of two consecutive screws using the hollow mill, we saw a crack in the bone between two holes. We used a long diaphyseal fitting stem to bypass the stress riser.

Follow‐up and Evaluation

All surgeries were performed by a single surgeon. Average follow‐up was 5 years (range, 2–8 years). Postoperatively, 15 patients began with strict non‐weight bearing walking for a period of 4 weeks, following which patients started on partial weight bearing with the use of a walker based on osteointegration on standard radiographs.

Postoperative outcome and patient assessment were done clinically and radiologically. Clinical assessment was done using the Harris hip score. Standard antero‐posterior radiographs of the pelvis were done preoperatively, postoperatively (on the day of surgery), 4 weeks postop, 3 months postop, 6 months postop, 12 months postop, and then at every 12‐month interval. For cemented arthroplasty, we used radiological criteria to assess lysis according to Gruen's zones6. For uncemented arthroplasty, the criteria proposed by Engh et al. were used7. Heterotopic ossification was graded according to Brooker's grading8.

Statistical Description

Students t‐test was used for statistical analysis. We used SPSS 17.0 software (SPSS Inc., Chicago, US) for analyzing all the statistical data and P value < 0.05 was considered significant for all the parameters.

Results

Functional Results

The mean Harris hip score improved by 155.34% from a mean of 35 ± 7.975 before the surgery to 89.38 ± 4.870 at the time of the latest follow‐up (P < 0.001). Improvement was significant for patients’ perceived pain, function, and mobility as assessed using the Harris hip score questionnaire. Pain improved by 79% after the surgery (P = 0.01). Patients’ function improved by 92% (P = 0.016). The percentage of improvement was significant for mobility among all three parameters, and it had improved by 171% (P = 0.006). At the latest follow‐up, 11 patients (68.75%) had no or slight and occasional pain without compromise in activity.

Radiological Results

There was no osteolysis around any acetabular component. Stem sinking and type A_L (Vancouver classification for periprosthetic fractures) periprosthetic fracture was seen in one case. For the radiographic evaluation, there was no significant difference in cup abduction angle at 3 months and at the latest follow‐up, indicating no implant migration (41.7° ± 5.6° vs 42.06° ± 6.1°, respectively, P = 0.729). Fourteen acetabular reconstructions (87.5%) presented no radiolucent line around the cup and 2 (12.5%) presented a non‐progressive and less than 1 mm radiolucent line in zone III.

Complications

A 50‐year‐old obese women who developed a type A_L periprosthetic fracture due to self‐fall was treated conservatively and ambulated non‐weight bearing for a period of 2 months. She started partial weight bearing with a walker 3 months after the surgery based on the radiographic appearance of callus formation. At 6 months and 12 months after the surgery, all patients were able to carry out daily activities like walking and climbing up and down a flight of stairs.

Discussion

The published literature does not mention the propensity of DHS to cause arthritis after the fracture has been adequately fixed and healed. We observed eccentric collapse of the head of the femur around the DHS. Conventional practice is to avoid removal of DHS unless it is complicated by infection or implant failure. There are no studies of THA have been done in a well‐fixed trochanteric fracture that has developed secondary osteoarthritis.

During preparation of the femur using the box osteotome, most of the lateral part of the neck is very thick compared to the greater trochanter. The direction in which the box osteotome is used is very important for preventing fracture of the greater trochanter as it is held by a reduced thickness of bone and the lateral cortex will have a huge hole from the barrel plate.

Hammad et al. reviewed 32 patients undergoing THA after failed DHS fixation and found only 1 periprosthetic fracture and 1 dislocation. Clinical results were good or excellent in 78% of patients at last follow‐up9.

Pachore et al. reviewed 30 patients who were followed up for an average of 20 months. They noted that a total of 12 patients (40%) were walking with the help of a cane, 3 (10%) with the help of a walker and 15 patients (50%) were walking without any aids. Of 30 patients, 16 (53.3%) had trochanteric non‐union preoperatively. Of the 16 patients, 4 patients had nonunion (fibrous) post‐arthroplasty. Harris hip score increased from a preoperative mean of 27.9 to a postoperative mean of 70.6 (range, 53–88). These patients required the use of a revision type femoral component in approximately one‐third of cases. These complications included a 5% rate of dislocation and a 4% rate of periprosthetic fracture. Periprosthetic fracture appears to be of greatest concern in patients with prior internal fixation of an intertrochanteric fracture10.

In our study, 1 patient developed a type A_L (lesser trochanter) periprosthetic fracture due to sinking CLS stem 3 months after surgery. She was a 50‐year‐old obese woman and was treated with a conservative method involving non‐weight bearing for 2 months. Patient was doing well after 2 years of follow‐up (Fig. 3).

Figure 3.

Patient was a 50‐year‐old obese woman (BMI = 30.11). The figure shows a healed Type A_L periprosthetic fracture, and the patient recovered after 2 months of strict non‐weight bearing walking with a height adjustable walker.

There have been several technical issues in converting healed fractures with proximal femoral implants into hip arthroplasty, such as extraction of implants, bone deformity, bone loss, poor bone quality, and associated trochanteric nonunion11, 12, 13. Revision of internal fixation for non‐united fractures has been reported to have good results in younger patients with good bone stock5. However, this is not the case with older patients, who formed the majority in our study, because older patients have poorer bone quality. This made implant removal and implant fixation a challenge as these patients were more susceptible to iatrogenic fractures despite taking all the necessary precautions.

As the weight bearing is over the calcar femorale in the tapered stems and because of the holes from removed screws, this area would have been weakened. There is a high chance of sinking of the stem and occurrence of periprosthetic fracture. In cases of bone loss over the lesser trochanter or when a mere 1–2 mm of bone remains over the lesser trochanter, it is better to use cylindrical fluted diaphyseal fitting stems than primary tapered stems. Thakur et al. evaluated 15 elderly patients treated with a tapered, fluted, modular, distally fixing cementless stem. At an average follow‐up of 2.86 years, the mean Harris hip score had improved from 35.90 preoperatively to 83.01 (P < 0.01)3. In our study, a Wagner SL revision stem of 190 mm length was used for 9 cases. We made them ambulate with full weight bearing from the first postoperative day with the help of a height adjustable walker.

Our study illustrates a novel technique that has not been described or discussed previously in the published literature. This technique can be used by surgeons when they have to remove well fixed DHS without damaging the neck. This is of particular use to surgeons who normally use the posterior approach for hip replacement surgery. It is not rational to use the direct anterior approach for THA in a case of well‐fixed DHS with hip arthritis14, 15. This technique may also be appropriate for a direct lateral approach after dislocating the femur anteriorly.

Limitations of this study are that the sample size is small. This is probably because over the past decade the use of a proximal femoral nails has superseded the use of DHS for fixation of inter‐trochanteric femur fractures. There are several articles on the failure of intertrochanteric fracture fixation and arthroplasty as a treatment option. However, the primary fractures of the patients in our series had healed and patients had developed secondary degenerative arthritis of the hip joint. Because the aim of our article was to describe the surgical technique rather than to measure the short‐term or long‐term outcomes, the sample size is sufficient.