Comparison of Clinical Outcomes of Periprosthetic Fractures Between Cemented vs Cementless Stem Methods After Initial Total Hip Arthroplasty or Bipolar Hemiarthroplasty: A Multicenter Analysis (TRON Study)

Abstract

Background

Periprosthetic femoral fractures are challenging complications after total hip arthroplasty (THA) or bipolar hemiarthroplasty (BHA). This multicenter study compared clinical outcomes of surgical periprosthetic femoral fracture (PFF) treatment following initial implantation of cemented vs cementless femoral stems.

Methods

This retrospective multicenter study included 201 patients from multiple hospitals who had PFFs treated surgically between 2010 and 2022. They were divided into THA (n = 71) and BHA (n = 130) cohorts based on initial arthroplasty. Clinical outcomes of mortality, radiographic bone union, operative metrics, and complications were compared after the patients were matched for age.

Results

In the matched THA cohorts (n = 15 each), there were no significant differences in mortality, bone union, operative time, blood loss, or complications between cemented and cementless stems. In the matched BHA cohorts (n = 27 each), mortality and bone union were similar, but cementless stems showed higher intraoperative blood loss. Furthermore, Type B fractures were more common with cementless stems, whereas Type C fractures were predominated with cemented stems.

Conclusions

Among patients who underwent THA or BHA, mortality and bone union following treatment of PFFs did not differ between cemented and cementless stems. However, patients who underwent BHA with cementless stems required longer operations and lost more blood. Type B fractures requiring revision occurred more frequently with cementless stems, whereas Type C fractures treated with open reduction and internal fixation were more frequent with cemented stems, which thus influenced perioperative factors. Cement use for initial arthroplasty was associated with certain favorable characteristics and positive treatment outcomes of PFFs.

Introduction

Total hip arthroplasty (THA) and bipolar hemiarthroplasty (BHA) have significantly improved the quality of life of patients suffering from idiopathic osteonecrosis of the femoral head, osteoarthritis, and fractures [[1], [2], [3], [4]]. As the demand for THA and BHA continues to rise, so does the incidence of periprosthetic complications, including periprosthetic femoral fractures (PFFs). According to the UK National Joint Registry of 2022, the incidence of revision arthroplasty for PFF has increased, and 44% of single-stage revisions are due to PFFs. Next to aseptic loosening, periprosthetic fracture is the second most common indication for revision hip arthroplasty [5].

PFF is a challenging complication that imposes substantial clinical and economic burdens on society because the management of PFF is technically demanding and associated with high complication rate, mortality rate, and treatment cost [6,7]. Recent data indicate that the 1-year mortality rate following surgery for PFFs is 11.2%, and older age, poor activities of daily living, and longer surgery time are related to higher 1-year mortality rate [8]. In a previous literature review, various risk factors for developing PFFs were identified, such as advanced age, female gender, post-traumatic osteoarthritis, osteoporosis, rheumatoid arthritis, low-energy trauma, and revision THA [9].

Fixation techniques for hip arthroplasty include cemented and cementless constructs. Although the use of cemented stems is advocated for preventing PFFs, fractures around cemented stems may yield inferior outcomes compared to those around cementless stems [10]. Thus, implant type (cemented or cementless stem) is another important risk factor associated with PFFs [7,11]. However, BHA with cemented stems is recommended for fragility fractures in elderly patients because of lower risk of PFFs compared to cementless stem. According to 1 study, rates of implant-related complications were significantly lower in the group treated with a cemented implant, and trends toward better function and better mobility in the cemented group were observed [12]. Several large-scale studies have shown that cemented stems have a lower incidence of PFF compared to cementless stems [[12], [13], [14]]. Nevertheless, despite the risk reduction associated with cemented stems, the presence of cement in cases of PFFs may impede bone healing and complicate treatment [15]. Additionally, the necessity of cement removal during revision surgeries can escalate the difficulty of managing PFFs [16]. There are few studies or multicenter analyses that directly compare how the choice of fixation method, cemented or cementless at the initial surgery, affects the postoperative outcomes of PFFs.

In this study, we aimed to compare periprosthetic fractures around cemented stems and cementless stems across patients in multiple hospitals, with the goal of identifying factors that influence treatment approaches and differences in the clinical outcomes.

Material and methods

Subjects

This multicenter retrospective study was approved by the ethics commission at each participating hospital. All patients provided their informed consent to participate in the study. Hospitals of the Trauma Research Group (TRON) have registered orthopaedic trauma surgery cases in the TRON database annually since 2010. The hospitals participating in the database are all associated with the Department of Orthopedic Surgery of our university, and orthopaedic surgeons perform surgical operations at these hospitals located in Central Japan. We collected cases of PFFs treated surgically from this database.

From 2010 to 2022, 256 patients who suffered PFFs were treated in all hospitals. After excluding 55 patients who were treated conservatively, we included the remaining 201 patients who had PFFs treated surgically. We divided these patients into two groups according to their primary arthroplasty: 71 patients received THA and 130 patients received BHA (Fig. 1).

Figure 1.

Patient flowchart.

Data collection

The following data were collected from electronic medical records: patient age at time of fracture, sex, body mass index, American Society of Anesthesiologists classification [17], residence before injury (own home, nursing home, or hospital) [8,18], preinjury mobility level (graded as independent, cane ambulation, walker or frame, wheelchair, and bedridden) [19], primary implant cemented or not, and date of death or date of the last follow-up. We also recorded the history of osteoporosis treatment and the presence or absence of previous fragility fractures, which were defined as fractures of the following bones occurring before the actual PFF: vertebra, proximal femur, distal radius, and proximal humerus.

Additionally, patient comorbidities present at the time of injury were collected from electronic medical records to allow for the calculation of the Charlson Comorbidity Index [20]. This index is calculated by assigning 1 point each to myocardial infarction, congestive heart failure, peripheral vascular disease, dementia, chronic pulmonary disease, connective tissue disease, peptic ulcer disease, mild liver disease, and diabetes. Two points each are assigned to hemiplegia, moderate to severe renal disease, complicated diabetes, malignancy within 5 years of the diagnosis, leukemia, or lymphoma. Three points are assigned to moderate to severe liver disease, and 6 points each are assigned to AIDS (not HIV infection) and metastatic solid tumor. We also recorded the Parker Mobility Score before injury [21].

Bone union was evaluated using the modified Radiographic Union Score for Tibial fractures (mRUST score), which ranges from 0 to 4 points [22].

We also extracted information associated with the surgical procedure, including operation time, blood loss during the operation, waiting time from initial visit to surgery, postoperative blood transfusion, and type of fixation.

Data analyses

To compare the outcomes between the groups, a matched-pairs analysis was performed to control for potential confounding variables, particularly age, because we found a statistically significant difference in the age of the patients who had PFF depending on the use of a cemented or cementless stem. Matching by age was deemed essential to minimize the influence of age-related factors on the outcome measures and to enhance the comparability of the groups.

Histograms and the Shapiro–Wilk test were used to verify whether the data followed a normal distribution. Independent samples that were normally distributed were evaluated with the Student t-test, and those that were not normally distributed were evaluated with the Mann–Whitney U test. Categorical variables were compared using Fisher’s exact test.

We estimated mortality after surgery for PFFs using Kaplan–Meier curves and compared differences in survival with a log-rank test. A P value <.05 was considered to indicate statistical significance. All statistical analyses were performed using EZR (version 1.5; Jichi Medical School, Saitama, Japan).

Results

THA cohort

After matching, 15 cases each were extracted. The baseline characteristics of the patients who received THA with a cementless or cemented stem are shown in Table 1.

Table 1.

Patients demographics of THA (n = 30).

Variable	Cementless	Cement	P value
Patients, n	15	15
Mean age, y (SD)	77.73 (7.31)	77.53 (7.24)	.94
Sex, n (%)
Male	3 (20)	4 (26.7)	1.00
Female	12 (80)	11 (73.3)
Mean BMI, kg/m3 (SD)	22.87 (4.86)	22.27 (3.71)	.71
ASA-PS
Ⅰ	5 (33.3)	4 (26.7)	.67
Ⅱ	8 (53.3)	10 (66.7)
Ⅲ	2 (13.3)	1 (6.7)
Preinjury mobility, n (%)
Independent	6 (40)	6 (40)	1.00
One cane	7 (46.7)	7 (46.7)
Walker or frame	1 (6.7)	1 (6.7)
Wheelchair	1 (6.7)	1 (6.7)
Bedridden	0 (0)	0 (0)
Mean PMS (SD)	6.33 (2.06)	7.07 (1.71)	.30
CCI, n (%)			.81
0	7 (46.7)	6 (40.0)
1	5 (33.3)	4 (26.7)
≥2	3 (11.1)	5 (18.5)
Residential status, n (%)			.60
Own home	14 (93.3)	12 (80.0)
Nursing home	1 (6.7)	2 (13.3)
Hospital	0 (0.0)	1 (6.7)
Treatment, n (%)			.71
ORIF	10 (66.7)	11 (73.3)
Revision	5 (33.3)	4 (26.7)
Vancouver classification			.161
A	1 (6.7)	0 (0)
B1	4 (26.7)	2 (13.3)
B2	4 (26.7)	1 (6.7)
B3	1 (6.7)	2 (13.3)
C	4 (26.7)	10 (66.7)
Median follow-up period (mo)	28.0 (1-72)	35.0 (0-140)	.33
Implants used in revision surgery
Exeter (Stryker)	2	1
Delta-Lock (TEIJIN)	0	2
Wagnar SL (Zimmer Biomet)	2	1
Profemur (Wright Medical)	1	0

ASA-PS, American Society of Anesthesiologists-Physical Status; BMI, body mass index; CCI, Charlson Comorbidity Index; ORIF, open reduction internal fixation; PMS, Parker Mobility Score; SD, standard deviation.

THA was performed in all patients who had osteoarthritis or avascular necrosis of the femoral head. There were no statistically significant differences in the patients’ background. However, there was a tendency for Vancouver Type C fractures to be more common with cemented stems. There was no significant difference in mortality rate (P = .92). The Kaplan–Meier plots of overall survival after surgery for PFFs are shown in Figure 2. The 1-year survival rates were 0.93 for cemented stems and 0.93 for cementless stems. There was no significant difference in overall survival between the 2 groups (P = .92). Operative time, intraoperative blood loss, mRUST score, and postoperative complications are shown in Table 2. There were 3 cases of reoperation in the cemented group. One patient underwent stem replacement due to loosening that occurred after open reduction and internal fixation (ORIF). One patient involved irrigation and debridement due to deep infection. Another case involved nonunion after ORIF, which required additional plating and autogenous bone grafting. There were no statistically significant differences between THA with cementless or cemented stems for any of these factors.

Figure 2.

Kaplan–Meier curves with patients’ death as the endpoint for the THA cohort, comparing cemented vs cementless techniques (after matching).

Table 2.

THA Operative time, blood loss, and postoperative complications (n = 30).

Variable	Cementless (n = 15)	Cement (n = 15)	P value
Median blood loss (range, ml)	432 (55-930)	420 (0-1950)	.58
Mean operative time (range, min; SD)	153.4 (66-289;56.0)	188.9 (66-461;103.2)	.25
Complications (%)
DVT	0 (0)	0 (0)	NA
PE	0 (0)	0 (0)	NA
AMI	1 (6.7)	0 (0)	1.00
Superficial infection	0 (0)	0 (0)	NA
Deep infection	0 (0)	1 (6.7)	1.00
Stroke	0 (0)	0 (0)	1.00
UTI	3 (20.0)	2 (13.3)	1.00
Pneumonia	1 (6.7)	1 (6.7)	1.00
Dislocation	0 (0)	0 (0)	NA
Loosening	0 (0)	1 (0)	1.00
Refracture	0 (0)	1 (0)	1.00
Reoperation	0 (0)	3 (20.0)	.22
Median mRUST score
6 mo	1 (0-2)	2 (0-2)	.55
12 mo	2 (0-2)	2 (0-2)	.95
24 mo	2 (0-2)	2 (2-2)	.11

mRUST, Modified Radiographic Union Scale for Tibia; DVT, deep vein thrombosis; PE, pulmonary embolism; AMI, acute myocardial infraction; UTI, urinary tract infection.

BHA cohort

After matching, 27 cases each were extracted. The baseline characteristics of the patients who underwent BHA with a cementless or cemented stem are shown in Table 3.

Table 3.

Patients demographics of BHA (n = 54).

Variable	Cementless	Cement	P value
Patients, n	27	27
Mean age, y (SD)	81.89 (7.95)	81.89 (7.94)	1.00
Sex, n (%)
Male	9 (33.3)	6 (22.2)	.544
Female	18 (66.7)	21 (77.8)
Mean BMI, kg/m3 (SD)	19.74 (3.47)	20.23 (2.59)	.575
ASA, n (%)
Ⅰ	6 (22)	4 (14.8)	.763
Ⅱ	13 (48.1)	16 (59.3)
Ⅲ	8 (29.6)	7 (25.9)
Preinjury mobility, n (%)
Independent	7 (25.9)	10 (37.0)	.59
One cane	7 (25.9)	9 (33.3)
Walker or frame	8 (29.6)	4 (14.8)
wheelchair	4 (14.8)	4 (14.8)
Bedridden	1 (3.7)	0 (0)
Mean PMS (SD)	5.33 (2.69)	5.52 (2.79)	.81
CCI, n (%)
0	6 (22.2)	8 (29.6)	.90
1	10 (37.0)	8 (29.6)
≥2	11 (40.7)	11 (40.7)
Residential status, n (%)			.33
Own home	17 (63.0)	22 (81.5)
Nursing home	7 (25.9)	4 (14.8)
Hospital	3 (11.1)	1 (3.7)
Vancouver classification			.0017
A	0 (0)	0 (0)
B1	11 (40.7)	3 (11.1)
B2	10 (37.0)	4 (14.8)
B3	1 (3.7)	2 (7.4)
C	5 (18.5)	17 (63.0)
Treatment, n (%)			.039
ORIF	18 (66.7)	25 (92.6)
Revision	9 (33.3)	2 (7.4)
Mean follow-up period (mo)	28.0 (0-95)	24.0 (0-145)	.917
Implants used in revision surgery
Exeter (Stryker)	4	1
Delta-Lock (TEIJIN)	0	1
CPT (Zimmer Biomet)	1	0
Wagner SL (Zimmer Biomet)	2	0
Versys (Zimmer Biomet)	2	0

ASA-PS, American Society of Anesthesiologists-Physical Status; BMI, body mass index; CCI, Charlson Comorbidity Index; PMS, Parker Mobility Score; SD, standard deviation.

BHA was performed in all patients who had femoral neck fractures. There was no statistically significant difference in the patients’ background other than for treatment methods (ORIF or revision) and Vancouver classification. There was a higher prevalence of Type B fractures with cementless stems, whereas a tendency for Type C fractures was more common for cemented stems. However, the difference in mortality rate was not significant (P = .77). The Kaplan–Meier plots of overall survival after surgery for PFFs are shown in Figure 3. The 1-year survival rates were 0.82 for cemented stems and 0.83 for cementless stems, with no significant difference in overall survival indicated between the 2 groups (P = .74). Data on operative time, intraoperative blood loss, mRUST score, and postoperative complications are shown in Table 4. Three cases of reoperation were performed in the cementless group and 2 cases in the cemented group. Among the 3 cases in the cementless group, 1 patient underwent conversion to a cemented stem due to loosening following ORIF. Two cases required implant removal due to infection, and in 1 of these cases, infection control was not achieved due to lower limb ischemia caused by diabetes, ultimately necessitating hip disarticulation. In the cemented group, 1 case required plate replacement due to a refracture, while the other case involved additional screw fixation at the femoral condyle following a fall-induced fracture. In the patients who underwent BHA, surgery for PFF of cementless stems tended to be associated with more intraoperative blood loss than that for cemented stems; however, there were no significant differences in mRUST and postoperative complications.

Figure 3.

Kaplan–Meier curves with patients’ death as the endpoint for the BHA cohort, comparing cemented vs cementless techniques (after matching).

Table 4.

BHA operative time, blood loss, and postoperative complications (n = 54).

Variable	Cementless (n = 27)	Cement (n = 27)	P value
Median blood loss (ml)	508.0 (100-3042)	229.5 (1-3900)	.018a
Mean operative time (min; SD)	157.04 (52-361;62.5)	167.6 (9-481;90.3)	.62
Complications (%)
DVT	2 (7.4)	1 (3.7)	1.00
PE	0 (0)	0 (0)	NA
AMI	0 (0)	1 (3.7)	1.00
Superficial infection	1 (3.7)	1 (3.7)
Deep infection	2 (7.4)	0 (0)	.49
Stroke	2 (7.4)	0 (0)	.49
UTI	3 (11.1)	2 (7.4)	1.00
Pneumonia	3 (11.1)	0 (0)	.24
Dislocation	0 (0)	0 (0)	NA
Loosening	1 (3.7)	0 (0)	1.00
Refracture	0 (0)	2 (7.4)	.49
Reoperation	3 (11.1)	2 (7.4)	1.00
Median mRUST score
6 mo	1 (1-3)	1 (0-2)	.71
12 mo	2 (1-4)	2 (0-2)	.08a
24 mo	2 (1-4)	2 (1-2)	.25

mRUST, Modified Radiographic Union Scale for Tibia; DVT, deep vein thrombosis; PE, Pulmonary embolism; AMI, Acute myocardial infraction; UTI, Urinary tract infection; SD, standard deviation.

^aP < .05.

Discussion

In this multicenter study, mortality and mRUST scores did not significantly differ between cementless and cemented stems in either the THA or BHA cohort. However, in the BHA cohort, operative time and intraoperative blood loss tended to be higher with revision of periprosthetic femoral fractures around cementless stems compared to cemented stems. Type B fractures were more prevalent with cementless stems, whereas Type C fractures were more common with cemented stems. Thus, cemented stems might not only reduce the occurrence of periprosthetic fractures but also yield favorable treatment outcomes and prognosis even in cases of fracture occurrence.

The incidence of PFFs is increasing due to the increasing number of primary THAs being performed. The incidence of periprosthetic proximal femur fractures per 100,000 person-years was 5.3 ± 2.2 (95% confidence interval (CI)) in 2004 and increased to 8.9 (95% CI) in 2016. The incidence of periprosthetic distal femur fractures per 100,000 person-years was 3.4 ± 1.7 (95% CI) in 2004 and increased to 4.4 (95% CI) in 2016 [23]. The treatment for PFFs commonly depends on various factors, including type of fracture, stability of the implant, quality of the bone stock, and patient age, comorbidities, and functional demands. Surgical treatment may require not only ORIF but also revision surgery or bone grafting [24].

While Vancouver A, B1, and C PFFs can be treated with ORIF, revision surgery is the standard treatment for Vancouver B2 and B3 PFFs [23,25,26]. However, these treatment algorithms do not take into account essential factors such as patient’s age, activity, and comorbidities. Generally, revision surgery is more invasive than ORIF [25,27]. In elderly, inactive, and frail patients with Vancouver B2 PFFs around well-cemented polished stems with an intact bone–cement interface, noninferior outcomes of ORIF were reported compared with revision [27]. There seems to be no consistent consensus of treatment methods between periprosthetic fractures around cemented and cementless stems, as shown in the present study. Treatment methods are often decided based on the Vancouver classification, but there were also cases where osteosynthesis was chosen for type B2 fractures, for which revision surgery is the first choice, and cases where revision surgery was performed for type C fractures, for which osteosynthesis is generally performed. In reality, it is thought that the decision is made based on the patient's background and the more detailed fracture type.

It has been previously reported that the type of implant, whether cemented or not, did not emerge as a significant risk factor for 1-year mortality after surgery for PFF. Instead, age and preinjury mobility level, particularly wheelchair status before injury, were identified as independent predictors of mortality outcomes [8]. Therefore, we performed a matching analysis based on age in the present study and concluded that there was no significant difference in mortality rates.

According to a study reporting about revision for periprosthetic femoral fracture, press fit stem may have lower mortality, lower estimated blood loss, similar subsidence, and hospital length of stays when compared to those with a cemented stem [10]. Thus, concerns have been raised that cemented stems might impact rates of bone union and clinical outcomes. However, the present study found that the presence of cement did not result in different rates of bone union or worse outcomes. The literature on the relationship between cement and bone union is limited. One study examining bone union in THA with femoral shortening osteotomy showed that cemented stems may impair bone union when inserted into the osteotomy site [15]. Our study focused on fractures occurring around a hardened cement mantle, which is different from the conditions examined in the previous research. Given this discrepancy, the presence of a hardened cement mantle may not be disadvantageous for bone union after fracture.

Furthermore, previous studies have indicated that revisions yield better outcomes for Type B2 and B3 fractures, which were more prevalent in cementless stems in the present study, compared to ORIF [28]. This suggests a tendency to choose revision over ORIF for Type B fractures. However, removal of the cement mantle has been associated with numerous complications and challenges [16,29]. Consequently, cemented stems tended to be managed with ORIF rather than revision. Moreover, the difference in blood loss may be attributed to the higher invasiveness typically associated with revision surgery [30], which involves additional procedures such as intramedullary canal management.

There was a significant predominance of Type C fractures with cemented stems in the patients who underwent BHA. Previous research has suggested an association between Type C fractures and osteoporosis [31]. While THA is typically performed for hip joint conditions such as osteoarthritis or femoral head necrosis, BHA is conducted for patients who have femoral neck fractures, which is indicative of bone fragility. The higher prevalence of osteoporosis in patients may contribute to these findings. In recent years, there has been a tendency to prefer the use of cemented stems in patients who have osteoporosis. This preference may also explain the higher prevalence of osteoporosis among patients receiving cemented stems. Additionally, Type C fractures treated with ORIF are of concern due to failure in transverse fractures [32], and some reports recommend additional fixation. However, favorable outcomes have also been reported [26]. Thus, the treatment approach for Type C fractures remains controversial and is likely influenced by patient characteristics and clinical presentation. In the present study, there was a significant predominance of ORIF with cemented stems in the patients receiving BHA. As mentioned earlier, revisions with cemented stems are known to be associated with a higher degree of invasiveness and complications. Given that patients undergoing BHA are generally older and have a lower physiological reserve, there appears to be a tendency toward selecting the less invasive ORIF procedures.

The present study was associated with several limitations. This is a retrospective study, and data collection was confined to information accessible in the medical records. Ensuring uniformity in the follow-up protocol across all participating hospitals in this multicenter retrospective study was not feasible. Additionally, treatment decisions were made by the attending surgeon based on the type of fracture and the patient’s health status, resulting in nonstandardized treatment approaches. The choice between a cemented or cementless stem may also have been influenced by the surgeon’s preference, which could have affected the results. In this study, cementless stems were used significantly more frequently, resulting in a considerable reduction in the cohort size after matching. These factors made the data heterogeneous and the matching resulted in a small sample size. As with the small sample size, this study only included patients from some of the related hospitals of Nagoya University, so there is a possibility of bias due to regional characteristics of treatment. Despite suggestions of an association between Vancouver Type C fractures and osteoporosis, this study did not include measurements of bone density. Furthermore, we did not evaluate the survivorship of stems after surgery for PFFs. Therefore, further research is needed to account for these limitations.

Conclusions

Periprosthetic fractures around cementless stems often involve Type B fractures, which frequently necessitate revision surgery. As a result, there tends to be more intraoperative blood loss compared to periprosthetic fractures around cement stems, which typically involve Type C fractures. Cemented stems may not only lower the frequency of periprosthetic fractures but also result in positive treatment outcomes and prognosis, even when fractures occur.

Availability of data and material

The datasets during and/or analyzed during the current study available from the corresponding author on reasonable request.

Ethic approval

The approval number in our institute is 2020-0564

Conflicts of interest

The authors declare there are no conflicts of interest.

For full disclosure statements refer to https://doi.org/10.1016/j.artd.2025.101699.

CRediT authorship contribution statement

Hiroaki Tachi: Writing – original draft. Yasuhiko Takegami: Writing – review & editing, Conceptualization. Toshiaki Okura: Writing – review & editing, Writing – original draft. Katsuhiro Tokutake: Writing – review & editing, Data curation. Hiroaki Nakashima: Writing – review & editing, Methodology. Kenichi Mishima: Writing – review & editing, Methodology, Investigation. Takehiro Kasai: Resources, Data curation. Shiro Imagama: Writing – review & editing, Supervision, Conceptualization.

Appendix

Supplemental Table 1.

Before matching patients demographics of THA (n = 71).

Variable	Cementless	Cement	P value
Patients, n	53	18
Mean age, y (SD)	77.62 (10.05)	80.22 (9.03)	.016
Sex, n (%)
Male	12 (22.6)	4 (22.2)	1.00
Female	41 (77.4)	14 (77.8)
Mean BMI, kg/m3 (SD)	22.72 (4.91)	21.57 (4.19)	.38
ASA-PS
Ⅰ	10 (18.9)	4 (22.3)	.49
Ⅱ	38 (71.7)	13 (72.2)
Ⅲ	5 (9.4)	1 (5.6)
Preinjury mobility, n (%)
Independent	23 (43.4)	6 (33.3)	1.00
One cane	23 (43.4)	8 (44.4)
Walker or frame	5 (9.4)	2 (11.1)
Wheelchair	2 (3.8)	2 (11.1)
Bedridden	0 (0)	0 (0)
Mean PMS (SD)	7.19 (1.83)	6.78 (1.83)	.41
CCI, n (%)			.89
0	19 (35.8)	7 (38.9)
1	15 (28.3)	4 (22.2)
≥2	18 (34.0)	7 (38.9)
Residential status, n (%)			.30
Own home	48 (90.6)	14 (77.7)
Nursing home	4 (7.5)	3 (16.7)
Hospital	1 (1.9)	1 (5.6)
Vancouver classification			.217
A	3	0
B1	19	4
B2	9	1
B3	3	2
C	18	11
Treatment, n (%)			.53
ORIF	42 (79.2)	13 (72.2)
Revision	11 (20.8)	5 (27.8)
Median follow-up period (mo)	31.0 (0-104)	21.5 (0-140)	.455

CCI, Charlson Comorbidity Index; PMS, Parker Mobility Score; SD, standard deviation.

Bolded P values indicate statistical significance (P > 0.05).

Supplemental Table 2.

Before matching patients demographics of BHA (n = 130).

Variable	Cementless	Cement	P value
Patients, n	102	28
Mean age, y	79.48 (8.58)	82.50 (8.44)	.10
Sex, n (%)
Male	36 (35.3)	6 (21.4)	.18
Female	66 (64.7)	22 (78.6)
Mean BMI, kg/m3 (SD)	20.15 (3.74)	20.01 (2.79)	.856
ASA-PS
Ⅰ	20 (19.6)	4 (14.3)	.65
Ⅱ	64 (62.7)	17 (60.7)
Ⅲ	18 (17.6)	7 (25.0)
Preinjury mobility, n (%)
Independent	35 (34.3)	10 (35.7)	1.00
One cane	32 (31.4)	9 (32.1)
Walker or frame	19 (18.6)	5 (17.9)
Wheelchair	15 (14.7)	4 (14.3)
Bedridden	1 (1.0)	0 (0)
Mean PMS (SD)	5.95 (2.57)	5.46 (2.76)	.39
CCI, n (%)
0	23 (23.0)	9 (32.1)
1	31 (31.0)	8 (28.6)
≥2	46 (46.0)	11 (39.3)
Residential status, n (%)			.713
Own home	70 (68.6)	22 (78.6)
Nursing home	24 (23.5)	5 (17.9)
Hospital	8 (7.8)	1 (3.6)
Vancouver classification			.0000099
A	3	0
B1	37	3
B2	40	4
B3	1	2
C	20	18
Treatment, n (%)			.014
ORIF	72 (70.6)	26 (92.9)
Revision	30 (29.4)	2 (7.1)
Median follow-up period (mo)	26.0 (0-107)	22.5 (0-145)	.91

CCI, Charlson Comorbidity Index; PMS, Parker Mobility Score; SD, standard deviation.

Bolded P values indicate statistical significance (P > 0.05).