1 Introduction

Total hip arthroplasty (THA) is notably one of the most successful operations in terms of cost effectiveness and quality of life gained. However, despite new developmental designs, implants have a finite lifespan with meta-analysis showing 58 % of patients with a THA will last more than 25 years.¹ Revision THA is set to rise over the next decade with a projected rise of 137 % between 2012 and 2030 in the UK.² Complications following THA revision are more pronounced than in primary hip arthroplasty, including infection (17.3 %), redo revision (15.8 %), venous thromboembolic disease (11.1 %), dislocation (5.43 %), pulmonary embolism (3.24 %) and death (2.11 %).³ Mortality following revision is well documented to be significantly higher in more elderly patients with higher American Society of Anaesthesiologists (ASA) physical status classification system.^3–5

To our knowledge there is no paper studying the risk of mortality following revision hip arthroplasty stratified by revision indication. Worldwide pooled registry data has shown the main indications for revision THA - aseptic loosening (55.2 %), followed by dislocation (11.8 %), septic loosening (7.5 %), periprosthetic fractures (6 %), wear (4.2 %), pain without other reason (3.7 %), and implant breakage (2.5 %).⁶

The aim of this study is to compare survival of patients undergoing revision THA for different indications to enable surgeons to have further information to guide decision making apart from age and ASA.

2 Methods

The study was designed as a retrospective case series. Ethical approval was granted locally by the hospitals ethical committee. All surgeries were performed by the senior author (RS) in the same coastal district general hospital in the UK. Interrogation of the hospitals hip arthroplasty database was conducted to identify THA revisions done between 2009 and 2016 with a minimum 2-year follow-up period. Operative technique was the same for all patients - the posterior approach was utilized and an extended trochanteric osteotomy was performed if needed. The same postoperative protocol was implemented in all cases using an enhanced recovery protocol including postoperative antibiotics, full weight bearing on day 1, low molecular weight heparin, anti-embolic stockings and analgesia.

For each patient, radiographs, case notes, outpatient clinic records and operative notes were examined. Besides collecting baseline demographics such as age and sex, other information collected included indication for revision, ASA physical status classification system, date of surgery and date of death.

Statistical analysis was carried out using the statistical package for the social sciences (SPSS). Kaplan Meir analysis was used to plot a survival curve using mortality as the end point. Mortality was further stratified by revision indication using log rank testing to determine if the indication for surgery conferred a difference in survival.

3 Results

Retrospective analysis of the hospitals hip registry database revealed 198 consecutive cases that had been done using a tapered modular uncemented stem in 183 patients. Patients undergoing bilateral hip revision surgeries were counted as two separate cases. The mean age of the cohort was 76 years and 64 % (n = 118) of the population consisted of women. Demographics of the highlighted cases are summarized in Table 1. Of the total number of revisions, 63 % (n = 125) of cases underwent cup and stem revision and 37 % (n = 73) underwent only stem revision. Indications for revision are enumerated in Fig. 1.

Close

Fig. 1 Indications for revision with Restoration stem.

Export to PPT

Mortality data was collected for patients undergoing surgery. A total of 60 patients died (67 % survivorship) during the follow up period. Kaplan-Meier analysis for mortality is shown in Fig. 2. The 5-year survival following revision surgery was 78 %. Log rank analysis comparing mortality rates per indication for revision showed that revision for aseptic loosening had the best survival and revision for hemiarthroplasty had the worst (Fig. 3). Differences in survival were highly significant (P = 0.000002).

Close

Fig. 2 Kaplan Meier curve for mortality following revision surgery.

Export to PPT

Close

Fig. 3 Log rank analysis comparing mortality rates per indication.

Export to PPT

4 Discussion

Revision THA is technically challenging with numerous anatomic and implant related hurdles to overcome. Furthermore, for the patient the surgery is physiologically demanding and often results in prolonged rehabilitation. With a growing need for revision THA, it is imperative that surgeons make holistic surgical decisions accounting for the surgical and medical needs of the patient.

There is a paucity of literature specifically investigating mortality following hip revision surgery. In this series of 198 cases, we have demonstrated a high mortality rate, the overall mortality rate during the study period was 33 % (n = 60) and the 5-year mortality rate was 22 %. This reflects the risk of major revision surgery. The majority of the patients operated on were ASA II, however this was closely followed by ASA III, who are high risk patients. This study did not stratify for ASA grade, here we have stratified mortality according to revision indication and have shown there to be different 5-year mortality rates for different revision scenarios - revision of hemiarthroplasty (51 %), periprosthetic fracture (28 %), infection (14 %) and aseptic loosening (13 %). Log rank calculations demonstrate a highly significant difference in mortality rates for the different surgical indications (P = 0.000002). Previous studies have shown high mortality rates following periprosthetic fractures of 9.8 % at 1 year.⁷ Revision for periprosthetic joint injection has mortality rates ranging from 25% to 45 %.^8,9 There is little literature regarding revision of hemiarthoplasty to THA with one study quoting major intra/post operative complications in 35.75 % of cases.¹⁰ No previous studies have directly compared the mortality rates for different revision indications.

There is a high rate of revisions performed for periprosthetic fractures in our study. One hypothesis for this is that the hospital where the study was conducted is located in a region which serves a high proportion of elderly patients who are more likely to have osteoporosis and recurrent falls, making them more susceptible to periprosthetic fractures. Although not formally measured, it was noted during the x-ray analysis that a large proportion of the periprosthetic fractures occurred in patients that had loose stems and had a number of interval x-rays prior to sustaining the periprosthetic fractures, suggesting that these patients were being either followed up or had problems prior to the fracture. This is in keeping with Swedish registry data which shows that 70 % of periprosthetic fractures were around loose stems and 23 % of these stems were known previously to be loose.¹¹ The high mortality associated with periprosthetic fractures poses the question of whether a patient known to have a loose stem should be revised earlier despite having little symptomology to prevent fracture late on, as the results suggest a significantly lower mortality in patients having revision for aseptic loosening than periprosthetic fracture. This corresponds with a study which found that periprosthetic fractures carried a similar mortality rate to neck of femur fractures.¹² It further reported that there was increased survival in patients that had revision as treatment compared to patients that underwent open reduction internal fixation.¹² Patients with periprosthetic fractures are often slow to mobilize postoperatively due to pain and sometimes subject to prohibited full weight bearing. This makes recovery very slow and therefore these patients are more prone to respiratory infections, skin necrosis, clots and other complications in the same way that patients with neck of femur fractures are susceptible. This group of patients therefore need to be closely medically managed postoperatively to reduce associated morbidities.

There was a very high mortality associated with revision of hemiarthroplasty with 5-year survival at 49 %. Even though not formally reported in the results, all of the revised hemiarthroplasties were Austin Moore prostheses which is an uncemented monobloc trauma stem. This prosthesis is no longer used in our hospital trust and is specifically no longer recommended by the National Institute for Health and Care Excellence (NICE).¹³ This is due to high rates of pain, loosening and poor mobility.¹⁴ As per NICE guidance we now use a cemented modular hemiarthroplasty prosthesis for neck of femur fractures which are known to have better survival rates.¹³ Neck of femur fractures are known to largely occur in a very frail, comorbid subset of patients. Surgical treatment for this carries a high mortality rate, therefore revision procedures on these patients will carry an even more pronounced mortality rate. Based on these study results it is proposed that revision for failed hemiarthoplasty should be reserved for patients who are physiologically able to cope with the surgical undertaking, otherwise conservative or other less invasive management should be considered. However, as mentioned, it is envisaged with modern hip trauma stems the need for revision will decline.

It could have been hypothesized that patients undergoing two stage procedures would have a higher mortality than those undergoing single stage procedures due to the repeated physiological insult from major surgery. However, interestingly between years 0–5 there was a better survival in patients undergoing revision for infection than aseptic loosening and after year 5, patients being revised for aseptic loosening have a better survival. In this study all patients undergoing revision for periprosthetic joint infections underwent 2 stage revisions. This is in keeping with another study showing that single stage procedures had a higher mortality than two stage procedures for infection.¹⁵ Here we suggest that patients with aseptic loosening are likely to be older as generally aseptic loosening occurs progressively over a long period of time. Therefore, there is likely to be a higher morbidity and mortality when undergoing revision at an older age. Conversely, periprosthetic joint infection is generally likely to occur earlier in patients than aseptic loosening. These patients are often younger and are more likely to survive major surgery. Further work into this area would be useful.

An ageing population and accumulation of chronic diseases results in a medically complex subset of patients. Major surgery in this group of patients leads to high rates of preoperative mortality.¹⁶ Furthermore, surgically trained specialty doctors often have difficulties managing the complex medical needs of patients post operatively. We advocate a multi-disciplinary approach to treating these patients with anaesthetic pre-assessment and optimization as well as allied special input from physiotherapy, dietetics and occupational therapists. Geriatric surgical liaison services are a relatively new specialty who can help pre and postoperatively medically manage elderly patients and have been shown to improve post operative outcome measures such as better pain control, reduced length of stay, decreased rates of pneumonia and pressure sores, earlier mobilization and degreased length of stay.^17,18

We recognize that there are limitations to the study. Firstly, this is a retrospective case series purely looking at mortality, it was outside the scope of this paper to investigate the performance of the stem or other clinical outcomes which are important to capture overall outcomes following revision. Secondly, there is no control group. This study does not provide any data on the mortality of patients if they were treated conservatively with no surgery. Thirdly, we did not specifically identify the cause of death in each of the patients, we have inferred the mortality is attributed to surgery as result of treating this subset of patients. An additional review of the 90-day mortality of the cohort may have been more informative, given that complications in the early post operative phase are more likely to be directly attributed to surgery.

5 Conclusion

When planning revision surgery, it of paramount importance that emphasis is given to the age, ASA and functional status of the patient. However, we also suggest that attention should also be given to the indication for revision as we have shown that this also can have an effect on the mortality. This study can be used as another helpful tool in guiding surgeon in the decision-making process and importantly in appropriately counselling patients. We recommend a multidisciplinary approach to be taken to address the complex medical and surgical needs of these patients.

Funding

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

Data statement

Data is available on request.

Informed consent

Informed consent waiver was obtained from ethical committee.

Institutional ethical committee approval

Approval was granted by the institutional review board.

CRediT authorship contribution statement

Gihan Jayasinghe: Writing – original draft, Conceptualization, Methodology, Formal analysis, Writing – review & editing. Rohit Kumar: Writing – review & editing, Visualization. Chris Buckle: Methodology, Investigation. Parthiban Vinayakam: Supervision, Project administration. Richard Slack: Supervision, Project administration.