1 Introduction

Cannulated hip screws (CHS) remain a popular choice of fixation for intracapsular neck of femur fractures (NOF#), especially when minimally or nondisplaced. For displaced fractures, arthroplasty is often chosen due to concerns over the increased risks of nonunion and avascular necrosis, although fixation may still be attempted in younger patients. Even in the nondisplaced fracture group some authors have advocated arthroplasty over fixation.^1–3 CHS fixation has been described as having high failure rates between 14 and 23 % between 6 and 12 months.^4–6 There have also been concerns reported over the outcomes of the subsequent salvage procedure of total hip arthroplasty (THA).^7,8 In one retrospective study,⁷ revision rates of 15 % have been reported when looking at salvage THA, with rates of instability and periprosthetic fracture being higher than would be expected from primary THA. When performing THA associated with concomitant removal of metalwork, there have also been reports of increased risk of infection (9 %) as well as 5 year mortality when compared to primary arthroplasty.⁹ Level 1 evidence also exists, demonstrating significantly higher risks of deep infection, early dislocation as well as periprosthetic fracture when compared to primary THA for intracapsular neck of femur fractures.¹⁰ Most of this evidence looks at a range of fixation options for fractured neck of femur, including dynamic hip screws, CHS, as well as intramedullary nails.

This paper aims to assess outcomes and complications of salvage arthroplasty following failed fixation with CHS for both displaced and nondisplaced fractures. The primary procedures were all performed in a UK Major Trauma Centre over a 13 year time period and revised in the same centre.

2 Methods

Data were collected retrospectively using the departmental database (Bluespier, Droitwich, UK) over a 14 year period from January 01, 2007 to March 31, 2021. All patients receiving CHS were included, then added to the study population if they subsequently underwent conversion to THA.

Data were collected on patient demographics, mechanism of injury, fracture classification, time between CHS and THA, reason for fixation failure, clinical outcome of THA, and any complications.

Our primary outcome measure was the Oxford Hip Score (OHS) at 1 year following total hip arthroplasty (THA) following failed CHS. We compared this to the OHS in a cohort of patients undergoing THA as a primary procedure for NOF# in the same local population. As secondary outcomes, we assessed complication rates in the salvage group, modes and times of failure of the initial fixation as well as degree of displacement of the original fracture.

3 Results

600 CHS were performed in 595 patients by varying grades of lead surgeon. To our knowledge, this represents one of the largest single reported cohort of CHS to date. There were 55 conversions to THA from this cohort, representing a 9.2 % salvage arthroplasty rate. Within this cohort, Oxford scores were available for 47 patients. Demographics and characteristics of this cohort can be seen in Table 1. Across the 55 THAs, there were 30 lead surgeons of varying grade. There were no significant differences in age or sex. ASA grade was shown to not have any significant impact on OHS at any time point. In this cohort of 55, there were 10 ASA 1 patients, 25 ASA 2's, 9 ASA 3's, 2 ASA 4's and 9 patients in whom a score was not recorded. At preop assessment for THA, there was a significant (p = 0.02) correlation between increasing age and worsening preoperative OHS, but this was not seen at any time point postoperatively. Mean follow up was 107 months (range 22–177, SD 50).

3.3

3.3 Patient reported outcomes (PROMs)

Patient-reported outcomes following salvage THA can be seen in Table 3 and Fig. 1. Overall mean (SD) OHS was 11.4 (8.0) preoperatively. This improved to 38.8 (10.4) at 1 year postoperatively. and 32.1 (14.9) at 5 years. This represents a clinically significant improvement, from severe impairment preoperatively to scores associated with mild to moderate osteoarthritis¹¹ maintained at 5 years. Between 2015 and 2020, 376 patients underwent primary THA for NOF#. The mean (SD) Oxford score was 39.7 (8.6) at 1 year and 39.4 (8.1) at 5 years postoperatively. No significant difference was found between implant subtypes, fixation method or bearing mobility option. Implant choice across the cohort included both single (n = 46) and dual mobility (n = 9), as well as fully cemented (n = 28), hybrid (n = 20) and uncemented (n = 7) fixation.

Table 3 Mean (SD) postoperative oxford hip scores classified by original fracture displacement.

Time Point	1 year ∗	5 years ∗
Displaced	44.1 (4.8)	40.5 (9.0)
Nondisplaced	35.2 (11.7)	21.4 (14.0)

Close

Fig. 1 Chart showing mean difference in postoperative oxford hip score at 1 and 5 Years in salvage arthroplasty cohort.

Export to PPT

4 Discussion

We have demonstrated that outcomes following salvage THA are equivalent to a primary THA performed for NOF#. A good outcome after THA for osteoarthritis when assessing PROMs has also been studied previously, where it was shown that an OHS of 38 at 12 months and 33 at 24 months was representative of a satisfied patient.¹³ Here, 2 year outcomes were not assessed, but at both 1 and 5 years, average OHS was within 1 point of these published figures. Other evidence reports an average OHS of 40–42 points being achieved at 1 year postop in cohorts of primary THA performed for osteoarthritis.^14,15 Our cohort's average OHS at 1 year of 38.8 represents an outcome which is within the minimum clinically important difference of 5 points when in the context of group analysis.¹²

In this cohort, the most common reason for failure was nonunion (44 %), followed by AVN (31 %). In a mixed cohort of displaced and nondisplaced original fractures, this is contrary to traditional concepts, as reported by Tsang et al. where it was found that AVN was more prevalent when looking at CHS failures.³ When looking at all 600 CHS procedures, the rate of patients who went on to have THA for AVN was 2.5 %. Other published evidence suggests a prevalence of around 8 %.¹⁶ There may be a proportion of patients who had AVN in our cohort following their CHS who did not then go on to THA, but this proportion is likely to be small. As such, the rate of AVN following CHS in our sample is also likely to be lower than other published data.

Complication rates following salvage also remained low in this cohort, with no cases of infection or instability at a minimum follow up of 2 years. Reported incidence of PJI following salvage arthroplasty ranges from 1.2 to 12 %^9,17–19 in cohort sizes similar to ours. Within these studies, other forms of internal fixation such as dynamic hip screws and intramedullary nails were included, which would require a more invasive removal procedure and potentially a higher rate of infection. Instability rates of around 5 % have been reported previously relating to salvage THA^19,20 which is of particular concern in a revision surgery setting in patients who often have poor quality soft tissues compared to a routine primary THA population. Our study demonstrates no cases of dislocation. Variation in implant choice can be seen in our results, representing an awareness of risks of instability in neck of femur fracture arthroplasty and a move towards the use of dual mobility implants. A variety of seniority level of lead surgeon can also be seen in our cohort, but in this centre it is local policy that each operating list is directly supervised by a consultant surgeon. Also, all THA performed for NOF# are done by either a senior trainee with direct lower limb arthroplasty trained consultant supervision or the lower limb arthroplasty consultant themselves.

This study is somewhat limited by the fact that long term follow up is not likely to be achieved to the same duration with NOF# patients compared to conventional THA outcome data obtained for primary osteoarthritis. As such, there are fewer responses in the 5 year follow up category, giving less statistical power. Longer term follow up than this is unlikely to be achieved due to the frailty and comorbidity associated with these patients leading to a higher mortality rate.

It has been demonstrated here that patients undergoing salvage THA following fixation for a displaced fracture did better than patients in whom the original fracture was nondisplaced. Confounding variables (age, sex, ASA grade) were also assessed and found not to have any significant influence. This is in contrast to other published evidence²¹ and raises the question as to whether CHS should be considered for more displaced fractures in suitable candidates.

Funding/conflicting interest declarations

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

The Authors declare that there is no conflict of interest.

Ethical statement

This work was performed in accordance with ethics principles of the Helsinki Declaration. Treatment was given in accordance to pre-existing and established protocols that involved no novel or experimental methods. Specific local ethics committee approval was not required based on the nature of the study undertaken, using data obtained with consent as part of normal clinical care and follow up of patients.

CRediT authorship contribution statement

Jack A. Turnbull: Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. Rupert Vicary-Watts: Data curation, Resources, Software, Supervision, Validation, Visualization, Writing – original draft. George J. McLauchlan: Conceptualization, Methodology, Project administration, Supervision, Validation, Writing – review & editing.