Predictive factor for bone cement implantation syndrome in osteoporotic femoral neck fracture undergoing cemented hip arthroplasty: a retrospective review of 142 cases

Background: Bone cement implantation syndrome (BCIS) is a serious and potentially fatal complication in cemented hip arthroplasty (CHA). Although previous studies have identied patients at risk for BCIS, few studies have reported on the association between BCIS and perioperative surgical factors in patients with osteoporotic femoral neck fracture (OFNF) undergoing CHA. This study aimed to identify the prevalence of BCIS and its related outcomes, correlating between BCIS and perioperative surgical factors in these patients. Methods: A single-centred, retrospective study was conducted between 2015 and 2018 with patients who sustained OFNF and underwent CHA. Perioperative data and BCIS-related outcomes were reviewed using electronic database and telephone interview. BCIS grading was classied according to the Donaldson classication. All patients were allocated into 2 groups: Group A (BCIS grade 0 or 1: BCIS-gr0/1) and Group B (BCIS grade 2 or 3: BCIS-gr2/3). Multiple logistic regression analysis was used to identify predictive factors for BCIS-gr2/3. Results: A total of 142 OFNF patients were included. The prevalence of BCIS-gr2/3 was 6.4% (n=9). Group B showed a signicantly higher major complication rate than Group A (44.4% vs. 8.3%, respectively, p<0.01). The in-hospital and 1-year mortality rates in Group B were also higher than in Group A, but non-signicantly different (p=0.12 and p=0.32, respectively). Multivariate analysis showed that preinjury non-ambulatory status (odds ratio [OR] 11.83; 95% condence interval [CI], 1.06–131.8, p=0.04), arrhythmia (OR 5.65; 95% CI, 1.13–28.93, p=0.04) and shape-closed femoral stem type (OR 9.78; 95% CI, 1.65–57.80, p=0.01) were the signicant independent predictors for BCIS-gr2/3 in these patients.

compromise, warfarin or diuretics use, American Society of Anesthesiologists (ASA) grades of 3 and 4, and use of long-stemmed femoral component-have been identi ed (1,9).
Despite the very high risk for BCIS, the use of bone cement in hip replacement surgery is still necessary to establish a well-xed prosthesis in the patients with poor bone quality, such as those with osteoporotic femoral neck fracture (OFNF). Therefore, to prevent BCIS during CHA for hip fracture, several recommendations-including thoroughly washing and drying the femoral canal before cement implantation, applying cement in retrograde fashion using a cement gun with a suction catheter and an intramedullary plug in the femoral shaft, and avoiding excessive pressurization-have been proposed (10). However, these recommendations did not include the femoral stem option that varies from a small taper-shaped (or "force-closed" type) to a large wedge-shaped (or "shape-closed" type) design. Moreover, previous studies had demonstrated that the stem pro le (11) and the large stem volume (12) are signi cantly associated with an increase of intramedullary pressure during stem insertion. Hence, the large wedge-shaped stem design should theoretically create higher intramedullary pressure and greater emboli formation during prosthesis insertion, resulting in higher risk for BCIS than with the small tapershaped femoral stem design. Moreover, to the best of our knowledge, only one study has addressed the risk factors for BCIS in patients with OFNF undergoing CHA (2). However, the previous study did not include some important surgical-related factors, such as time to surgery, operative time, type of operation, femoral stem type and intraoperative blood loss.
Therefore, the aim of the current study was (1) to investigate the prevalence of BCIS in patents with OFNF undergoing CHA and its impact on patient outcomes, and (2) to identify the risk factors for BCIS in these geriatric patients.

Study design, participants, and inclusion and exclusion criteria
The present study is a single-centred retrospective review that was approved by the Institutional Board Review (Protocol number ID 05-61-57). Hip fracture patients who underwent CHA during a 4-year period (between January 2015 and December 2018) were identi ed and reviewed from the hospital electronic records. The inclusion criteria were patients who had 1) sustained low energy trauma and been diagnosed with femoral neck fractures and 2) underwent cemented hip replacement. The exclusion criteria were patients who had 1) multiple fractures or poly traumatized injury, 2) revision surgery and 3) pathological fracture other than osteoporosis hip fracture, such as metastatic fracture.
Perioperative protocol, surgical procedure and implant selection All patients had received the same preoperative protocol with routine deep vein thrombosis (DVT) screening with doppler ultrasonography (13). In cases where patients had unexplainable dyspnoea with hypoxemia that was clinically suspicious for pulmonary embolism (PE), computer tomography angiography (CTA) was performed. In case of positive DVT nding, the patient would receive anti-coagulant injection for 2 weeks prior to de nitive operation and then continue as oral anti-coagulant for total 3-months duration. The anaesthetic technique, either general or regional anaesthesia, was based on the decision of the respective anaesthesiologists. The surgical approach was an anterolateral approach with anterior hemimyotomy of the gluteus medius, in the lateral decubitus position. The type of operation, either cemented bipolar hip arthroplasty (BHA) or hybrid total hip replacement (THA), and the type of femoral stem were selected according to the respective surgeons' preferences. For THA, the patients all received cementless acetabulum cup combined with cemented femoral stem. The cemented femoral stem choices in our institution included cemented Corail stem (DePuy International, Leeds, UK), C-stem (DePuy International, Leeds, UK), CPT stem (Zimmer Biomet, Warsaw, IN) and Exeter stem (Stryker Howmedica Osteonics, Mahwah, NJ). All polymethylmethacrylate (PMMA) bone cement used in this study was high-viscosity PALACOSâ bone cement (Heraeus Medical LLC, Yardley, PA). Our routine cementing technique was done in the following steps: 1) inform the anaesthesiologist for the cementation during the brushing and drying of the femoral canal, 2) manually mix the cement according to the manufacturer's instruction, 3) insert an intramedullary cement restrictor plug 2 cm below the tip of the femoral stem, 4) use the distal suction catheter on the top of the intramedullary plug and pull the catheter out as soon as it is blocked with cement, 5) insert cement with nger-packed technique, at 2 minutes after cement mixing, without proximal pressurization, and 6) insert femoral stem with distal centralizer (except for the cemented Corail stem which did not require distal centralizer) and wait for the hardening of cement.

Data collection and outcome measurement
The charts were reviewed, using the electronic hospital database, to obtain the required data, including demographic data, perioperative data and postoperative data. The following demographic data were collected: age, gender, body mass index (BMI), ASA physical status (14), comorbid disease, time to surgery, preinjury ambulatory status, anti-platelet and anti-coagulant medication use, and preoperative laboratory values, such as haemoglobin (Hb) level, platelet count, glomerular in ltration rate (GFR) and albumin level. The time to surgery de ned as the duration between the day of injury (fracture occurred) and the operative day. Preinjury ambulatory status was classi ed into 3 groups: ambulatory, ambulatory with assistance and non-ambulatory. Ambulatory was de ned as independent ambulation. Ambulatory with assistance was de ned as requiring assistance, such as with a cane or walker, for ambulation. Nonambulatory was de ned as a composite of wheelchair and bedridden status (15).
Perioperative data-including type of anaesthesia, operative time, estimated blood loss, intraoperative ndings related to BCIS and BCIS grading, and packed red cell (PRC) transfusion-were reviewed and recorded by one author (SP), an anaesthesiologist who was not involved in the data analysis. BCIS was de ned as sudden hypoxia, hypotension, cardiac arrhythmia and sudden cardiac arrest at the time of the cement insertion. BCIS grading was classi ed according to the classi cation system proposed by Donaldson (1) ( Table 1). Each patient was classi ed as having no BCIS (grade 0), or BCIS grade 1, 2 or 3, depending on the severity of hypotension, oxygen desaturation, degree of consciousness and cardiovascular collapse status. Femoral stem was classi ed in 2 types according to the classi cation proposed by Shen (16): shape-closed type (large stem geometry that fully occupies the medullary canal) and forced-closed type (small taper design that allows the subsidence of the femoral stem in the cement mantle). Therefore, in our study, the cemented Corail stem was classi ed as shape-closed type, while the C-stem, CPT stem and Exeter stem were all classi ed as force-closed type.
Postoperative data-including major perioperative complications, in-hospital death during admission, and the death within 1 year postoperatively-were collected via chart review and telephone interview. The major perioperative complications were de ned as follows: 1) cardiac complication requiring inotropic drug and admittance into coronary care unit, 2) pulmonary complication requiring ventilator support, 3) renal complication requiring peritoneal or haemodialysis, 4) infection requiring intravenous antibiotic treatment, and 5) symptomatic thromboembolic complication (acute stroke, symptomatic DVT or symptomatic pulmonary embolism).

Statistical analysis
MedCalc software version 15.8 was used to analyse the data. Normally distributed continuous data were presented as mean and standard deviation and compared with Student t-test, while non-normally distributed continuous data were presented as median and interquartile range. Categorical data were presented as proportion of cases and compared with Fisher's exact test or Chi-square test as appropriate.
All patients were allocated into 2 groups based on BCIS grading: Group A (BCIS grade 0 or 1: BCIS-Gr0/1) and Group B (BCIS grade 2 or 3: BCIS-Gr2/3). Risk factors for BCIS-Gr2/3 were compared between both groups. Univariate logistic regression analysis was used to evaluate the association between risk factors and BCIS-Gr2/3, and the predictive factors with values of p£0.10 were calculated by multivariate logistic regression analysis. Signi cance was de ned as values of p<0.05.

Results
General characteristic data of study population Between January 2015 and December 2018, 288 femoral neck fracture patients were treated in our hospital. Of these, 28 patients had been treated with multiple cannulated screws xation, whereas 260 patients had been treated with either cementless arthroplasty (104 patients) or cemented arthroplasty (156 patients). Fourteen patients who underwent CHA were excluded due to multiple fracture (n=1), revision surgery (n=4) and metastatic fracture (n=9). Therefore, a total of 142 OFNF patients were included in this study. Among those patients, the median age was 80.5 years (range 34-98 years), and 120 of them (84.5%) were female. Thirty-three patients (23.4%) were classi ed as ASA grade 4, and 70 patients (49.3%) were taking either anti-platelet or anti-coagulation medication prior to hospitalization.
The median time from injury to surgery was 4 days (range 1-180 days). There were nine patients having positive results from the preoperative DVT screening and they were all treated with anticoagulant injection for two weeks prior to surgery. Of these, one patient had experienced BCIS grade 1 while the other eight patients had experienced BCIS grade 0. BHA was used in 134 patients (94.4%), while 8 patients (5.6%) received THA. The force-closed type femoral stem was used in 94 patients (66.2%), while 48 of them (33.8%) received shape-closed type femoral stems. Regional anaesthesia was used in 114 patients (80.3%), whereas general anaesthesia was given to 28 patients (19.7%). The median operative time was 165 minutes (range 105-270 minutes). The median estimated blood loss was 300 mL (range 50-1650 mL), and the median total PRC transfusion was 1 unit (range 0-7 units).

Incidence of BCIS, perioperative complications and mortality rate
The incidence rates of BCIS grades 0, 1, 2 and 3 were 73.9% (n=105), 19.7% (n=28), 3.6% (n=5) and 2.8% (n=4), respectively. Table 2 shows the intraoperative ndings after cementation and the clinical information related to BCIS from 9 patients who had BCIS grade 2 or 3. Of these, four patients (44%) (no. 1, 3, 5, and 6) had experienced major perioperative cardiovascular complications (2 acute MI, 1 acute right-side heart failure, and 1 postoperative cardiac arrest), whereas the other ve patients (no. 2, 4, 7, 8, and 9) had been treated and recovered from BCIS without any major perioperative complication. The demographic data and perioperative data with simple statistical comparison between those who had BCIS-gr0/1 (Group A, n=133) and those who had BCIS-gr2/3 (Group B, n=9) are shown in Table 3. There was no signi cant difference in the demographic data for age; gender; BMI; ASA classi cation grade 4; pre-existing ischemic heart disease or congestive heart failure, renal failure, chronic obstructive pulmonary disease, stroke, cancer, and dementia; concomitant anti-platelet or anti-coagulant medications; preoperative laboratory values; type of anaesthesia; anaesthetic time; and type of operation between groups (p>0.05). However, compared to Group A, Group B had signi cantly higher incidences of preexisting cardiac arrhythmia, preinjury ambulatory status and shape-closed femoral stem type (p<0.05 all). Table 4 shows the in-hospital and 1-year postoperative mortality and major perioperative complications during admission. Postoperatively, two patients in each group died within the admission period. One patient (0.8%) in Group A died from respiratory failure at 4 weeks after the operation, and one patient (11.1%) in Group B died from cardiovascular failure 3 days after the operation. After discharge, four patients in Group A (3.8%) died from respiratory infection, acute myocardial infarction, acute stroke and gastrointestinal bleeding, respectively. None of the patients in Group B died after hospital discharge. The overall 1-year mortality rate was 4.2% (n=6), and the major perioperative complication rate was 10.6% (n=15). There was a statistically signi cant higher rate of major perioperative complications in Group B compared to Group A (44.4% vs. 8.3%, p<0.01). Group B also had non-signi cantly higher in-hospital and 1-year mortality rates than Group A (11.1% vs. 0.8%, p=0.12 and 11.1% vs. 3.8%, p=0.32, respectively).
Logistic regression analysis of the risk factors for BCIS

Discussion
BCIS is a serious and potentially lethal perioperative complication of cemented hip replacement which is theoretically caused by cement monomer mediation (17), immune reaction (18) or emboli during cement and prosthesis insertion (1,4). Among these hypotheses, the embolic model from the high intramedullary pressure during cement and prosthesis insertion is the most evidenced cause of BCIS (3,(19)(20)(21)(22). To the best of our knowledge, only few previous studies had been addressed the association between perioperative factors and BCIS. Moreover, a few studies also demonstrated that the association between the stem design and the increase of intramedullary pressure during stem insertion (11,12). Therefore, the aim of this study was to analyse the incidence and outcome of BCIS in OFNF patients undergoing cemented hip replacement, and to investigate the predictive factor for BCIS in these patients.
The incidence of overall BCIS in the present study was 26.1% (with BCIS grades 1, 2 and 3 as 19.7%, 3.6% and 2.8%, respectively), which is comparable with previous studies (2, 23). The in-hospital and postoperative 1-year mortality rates in patients with BCIS-gr2/3 were higher than those in patients with BCIS-gr0/1 (11.1% vs. 0.8%, and 11.1% vs. 3.8%, respectively), but did not have statistical signi cance. Although these results were dissimilar to the previous study by Olsen et al.
(2) that reported signi cantly higher 30-day and 1-year mortality in BCIS-gr2/3 compared with BCIS-gr0/1, this difference could be explained by the relatively small study population in the present study. Moreover, our results also showed that major postoperative complications in patients with BCIS-gr2/3 were also signi cantly higher than those in patients with BCIS-gr0/1 (44.4% vs. 8.3%, p<0.01). These results highlight the signi cance problem of this complication and emphasise the need for meticulous attention to preoperative and intraoperative management.
This study also revealed the signi cant correlation between BCIS-gr2/3 and the patient and surgical factors. From the multivariate regression analysis, our study showed that the independent factors for BCIS-gr2/3 were preinjury wheelchair or bedridden ambulatory status (OR=11.8, 95% CI 1.1-131.8, p=0.04), underlying cardiac arrhythmia (OR=5.7, 95% CI 1.1-28.9, p=0.04) and use of shape-closed femoral stem (OR=9.8, 95% CI 1.7-57.8, p=0.01). The poor preinjury ambulatory status and cardiac arrhythmia are both directly related to the patients' impaired cardio-pulmonary function and, therefore, result in signi cant physiological disturbance during surgery and cement implantation. This nding is also comparable to the previous studies that showed the signi cant association between BCIS-gr2/3 and patients' physiological factors, such as ASA physical status grade 3 or 4, underlying chronic obstructive pulmonary disease (2) and pulmonary hypertension (24).
The signi cant correlation between shape-closed femoral stem and BCIS-gr2/3 could be explained by the pressure-model mechanism as an effect of stem geometry. Previous study showed that the stem design as larger stem volume, use of proximal stem centralizer, and a lateral-medial tapered stem, signi cantly associated with increased cement pressurization and higher intramedullary pressure (12). Therefore, the shape-closed stem, that has larger proximal geometry than a force-closed stem, would result in higher intramedullary pressure during stem insertion (12) and release a greater number of embolic particles into the vascular system. Consequently, the shape-closed stem insertion should pose a higher risk of BCIS-gr2/3 than the force-closed stem.
The strengths of the present study are related to the preoperative DVT screening to exclude preoperative DVT in all cases and the use of the same surgical approach and cementation protocol in the treatment of osteoporotic femoral neck fracture in the elderly, as previously described. Therefore, some confounding factors-such as the type of bone cement used, the time of cement insertion and the cementing technique -could be controlled. Moreover, to the best of our knowledge, the present study is the rst study that demonstrates the impact of femoral geometry on the severity of BCIS. Hence, more studies on how the shape-closed femoral stem affects intramedullary pressure are required.
This study had the following limitations. First, regarding the retrospective nature of this study, some useful clinical information, such as smoking status and alcohol intake, was not available in all cases and, therefore, was absent from our analysis. Second, our sample size was relatively small and from only one centre. Thus, other possible risk factors-such as medications, anaesthetic technique and type of surgery -might remain undetected. Finally, the cementing technique used in our institution (use of high viscosity cement with manual mixing and nger-packed technique) did not comply with the previous recommendation (use of low viscosity cement with vacuum mixing and cement gun) (10), mainly due to the restriction of medical insurance in our country. This cementing technique protocol might affect the outcome of the present study. However, our results did not show a higher incidence of BCIS, perioperative complications or postoperative mortality compared with previous studies. Therefore, a well-designed multicentred prospective study with a larger sample size is recommended to explore the effect of other potential risk factors.

Conclusion
BCIS in OFNF patients undergoing CHA is common and associated with high morbidity and mortality rates. The results of this study showed that preinjury wheelchair or bedridden ambulatory status, underlying cardiac arrhythmia and the use of shape-closed femoral stem were signi cant independent factors for high-grade BCIS. To avoid BCIS, complete preoperative assessment and perioperative management are required. Patients with high risk of BCIS must be identi ed and individually managed following suitable guidelines for risk reduction. If possible, the shape-closed femoral stem should be avoided.

List Of Abbreviations
BCIS: Bone cement implantation syndrome