To our knowledge, this is the first consecutive cohort study to compare the relative clinical outcomes of cemented and uncemented hemiarthroplasty in the treatment of femoral neck fractures in patients with neuromuscular disease. This retrospective case-control study demonstrated that the uncemented hemiarthroplasty had shorter duration of surgery and less intraoperative blood loss, but the two procedures for patients with neuromuscular diseases were equally good regarding functional outcomes, health-related quality of life, mortality, or general and local complications.
There were several advantages in the use of uncemented implants, which were used in our study. The amount of duration of surgery and intraoperative blood loss can be reduced in uncemented implants compared with cemented implants, which will be partly due to the time for the polymerisation of the cement. Ng and Krishna recommend that uncemented hemiarthroplasty is preferred over cemented hemiarthroplasty because of reduced duration of surgery and intraoperative blood loss [14]. Figved and colleagues conducted a two-center randomized equivalence trial involving 230 patients in New Zealand and found the duration of surgery and intraoperative blood loss were less in the uncemented group [15]. Many of patients with neuromuscular diseases have complex deformities and extensive contractures, which makes them more challenging surgery [16], and as a result, the duration of surgery of these patients may be longer than that of patients without neuromuscular disease, and the intraoperative blood loss may be more. In addition, in our study, part of the reason for the increase of uncemented hemiarthroplasty in recent years may be that cemented fixation takes more time, and cement removal can be difficult if a revision surgery is required in the future. The shorter operative times and less intraoperative blood loss may have some organizational and economic benefits, but should not be overestimated.
Notably, none of the three pain and functional outcome scales, VAS, HHS, and EQ-5D, showed any differences between the groups in our study. Similar studies have also shown that in patients without neuromuscular disease, there were no differences in pain or patient-reported outcome measures at one year [17, 18]. Parker and colleagues found cement to be associated with less pain and better mobility and Inngul et al found better outcomes as measured by the Short Musculoskeletal Functional Assessment, HHS and EQ-5D scores for the cemented implant [5, 19]. However, our results suggested that cemented hemiarthroplasty cannot reduce the risk of pain and improve functional outcomes in patients with neuromuscular disease. As life expectancy increases for patients with neuromuscular disease, these patients will have higher expectations for mobility and quality of life. We found the cemented and uncemented implants used in our study were equally good in terms of the functional outcome and health-related quality of life.
Although the between-group difference was not statistically significant, we found that the uncemented group was associated with higher risk of postoperative periprosthetic fracture (5.9% in the cemented group vs. 1.0% in the uncemented group), which is supported by previous studies. Barenius and colleagues conducted a randomized trial comparing 67 cemented prostheses with 74 uncemented prostheses and found 6.8% periprosthetic fractures in the uncemented group, compared with 3% in the cemented group [20]. Similarly, Morris and colleagues found 5 periprosthetic fractures in the uncemented group (10.7 %) as compared to none in the cemented group [21]. The specific reason for the improved outcome observed in patients with cemented hemiarthroplasty has not been definitively clarified. Patients with neuromuscular diseases are well known to be at significant risk of osteoporosis [22], and imbalances in muscle strength may lead to an increased risk of falls, all of which may cause an increased incidence of periprosthetic fractures in such patients. One theory is that cemented fixation may be better resistant to periprosthetic fracture in patients with risk factors such as neuromuscular diseases, a history of falls, and osteoporosis [23]. Although our results show no significant difference in periprosthetic fractures between groups, we should be cautious about this complication. Large sample studies may be needed to assess whether the implants used in our study have a different risk of periprosthetic fracture.
Dislocation of a hemiarthroplasty is uncommon, with an incidence of 1.5–2.0% [24–26]. Evidences also suggested patients with neuromuscular diseases have a higher incidence of dislocation after hemiarthroplasty than those without, ranging from 4.8–45% [26–28]. The total incidence of dislocation in our study was 4.5%, slightly lower than in previous studies, with a dislocation rate of 7.8% in the uncemented group and 2.9% in the cemented group, with no significant difference between the two groups. Cognitive dysfunction from dementia, psychosis, or confusion is a reported risk factor for hip instability and neuromuscular dysfunction positively correlates with dislocation [29, 30]. However, Suh et al. found no difference in the incidence of dislocation between patients with or without neuromuscular disease by using a posterior soft tissue repair technique to maintain adequate soft tissue tension [28]. Some have suggested that the posterior approach is a risk factor for dislocation after hemiarthroplasty [31–33]31–33, but others have shown that surgical approach and dislocation do not correlate after hemiarthroplasty [26, 28, 34]. Because all of patients in our study were operated with posterolateral approach, comparison with another approach is needed. Based on our results and previous studies, patients with neuromuscular disease have a higher risk of dislocation than the general population, and both arthroplasties were appropriate.
In addition to periprosthetic fracture and dislocation, there are other complications such as pneumonia and deep vein thrombosis, which may be related to the fact that patients with neuromuscular diseases are less likely to follow the post-operative procedures and guidelines to facilitate rapid recovery without complications [35]. But there were no significant differences in these complications. Interestingly, the number of cardiovascular complications did not differ between the groups, contrary to the widely held view that there is an association between cement and cardiovascular complications [36].
Intraoperative mortality is the most worrying complication for patient, their family and surgeon. Intraoperative death almost exclusively occurred during cemented procedures, which may be caused by the BCIS [37], but no intraoperative death was found in our study. The Kaplan–Meier curve was performed to analyze patients’ survivorship and the result showed distinctly similarity among the two groups, which was comparable with previous studies. Fenelon et al found no difference in mortality between cemented and uncemented hemiarthroplasty at seven days, thirty days, and one year [38]. Similarly, we believe that the use of cement has no detrimental effect on the short- and mid-term mortality in patients with neuromuscular diseases.
Several limitations must be considered when interpreting the presented data. First, this was a retrospective study with limited patient numbers. Future high-quality studies with larger sample size and longer follow-up are warranted to confirm the results of our study. Second, patients did not undergo conventional dual-energy x-ray absorptiometry testing, so it is impossible to exclude the difference of bone mineral density, which could affect the risk of periprosthetic fracture. However, in our study, all patients developed a low-energy hip fracture, which met the clinical criteria for osteoporosis by definition. Third, some patients are dementia and cannot participate in the extensive follow-up, so we can only obtain the required follow-up score of such patients through their family members, which may have added a risk of bias. Since the scores were obtained after communicating with family members as fully as possible, the risk of bias is assumed to be limited.