Intramedullary Nailing with and without the Use of Bone Cement for Impending and Pathologic Fractures of the Humerus in Multiple Myeloma and Metastatic Disease

Simple Summary Although intramedullary nailing (IMN) is considered the standard of care for the surgical management of most femur metastatic diseases, the optimal treatment of metastatic humeral impending and/or pathologic fractures is still debatable. In this study, we explored the usage of cemented vs. uncemented IMN in treating both impending and pathologic fractures, secondary to metastatic disease or multiple myeloma, and compared the outcomes in terms of survival, function, blood loss, blood transfusions, and perioperative complications. Our findings demonstrated that both groups had comparable outcomes, except higher blood loss was found in the cemented group. Thus, intramedullary nailing, both with and without cement, is a relatively safe and effective therapeutic modality for metastatic humeral disease in select patients with similar clinical outcomes and acceptable complication rates. Abstract Although intramedullary nailing (IMN) is considered the standard of care for the surgical management of most femur metastatic diseases, the optimal treatment of metastatic humeral impending and/or pathologic fractures is still debatable. Moreover, the use of cemented humeral nails has not been thoroughly studied, and only a few small series have compared their results with uncemented nails. The purpose of this study was to compare the (1) survivorship, (2) functional outcomes, and (3) perioperative complications in patients receiving cemented versus uncemented humerus IMN for impending or complete pathologic fractures resulting from metastatic disease or multiple myeloma. We retrospectively reviewed 100 IMNs in 82 patients, of which 53 were cemented and 47 were uncemented. With a mean survival of 10 months (Cemented: 8.3 months vs. Uncemented: 11.6 months, p = 0.34), the mean Musculoskeletal Tumor Society (MSTS) scores increased from 42.4% preoperatively (Cemented: 40.2% vs. Uncemented: 66.7%, p = 0.01) to 89.2% at 3 months postoperatively (Cemented: 89.8% vs. Uncemented: 90.9%, p = 0.72) for the overall group (p < 0.001). Both cohorts yielded comparable complication rates (overall [22.6% vs. 19.1%)], surgical ([11.3% vs. 4.3%], and medical [13.2% vs. 14.9%], all p > 0.05), but estimated blood loss was significantly higher in the cemented group (203 mL vs. 126 mL, p = 0.003). Thus, intramedullary nailing, with and without cement augmentation in select patients, is a relatively safe and effective therapeutic modality for metastatic humeral disease with similar clinical outcomes and acceptable complication rates. While controlling for possible selection bias, larger-scale, higher-level studies are warranted to validate our results.


Introduction
The skeletal system represents the third most common site for metastases, following the lungs and liver, with the humerus being the second most affected long bone (after the femur), accounting for about 20% of all metastatic long bone lesions [1]. Bone metastases

Patient Selection and Demographics
The present study was a retrospective comparative analysis of a prospectively maintained institutional review board-approved single-surgeon database in an urban academic setting. Between August 2011 and July 2022, a total of 82 patients (41 [50%] males and 41 [50%] females) who underwent IMN for impending or complete pathologic humeral fractures, for metastatic disease (n = 28) or multiple myeloma (n = 54), were included. Some of these patients have been described, earlier, in a separate study that investigated the outcomes of single stage multiple nailing procedures [42]. Among these patients, 18  were 40 (40%) and 60 (60%) nails placed for impending and complete pathologic fractures, respectively ( Table 1). Out of 82 patients, 33 29 [80.6%]) had more than 1 nail placed in other long bones in 1 or more settings. The most common combination was unilateral humerus and unilateral femur IMN (18; 8 cement and 10 uncemented), followed by unilateral humerus and bilateral femora IMN (9; 2 cement and 7 uncemented). and 41 [50%] females) who underwent IMN for impending or complete pathologic humeral fractures, for metastatic disease (n = 28) or multiple myeloma (n = 54), were included. Some of these patients have been described, earlier, in a separate study that investigated the outcomes of single stage multiple nailing procedures [42]. Among these patients, 18 underwent bilateral IMN (8 cemented vs. 10 uncemented), bringing the total number of humeral nails placed to 100 (53 [53%] cemented and 47 [47%] uncemented) ( Figure 1). There were 40 (40%) and 60 (60%) nails placed for impending and complete pathologic fractures, respectively ( Table 1). Out of 82 patients, 33 29 [80.6%]) had more than 1 nail placed in other long bones in 1 or more settings. The most common combination was unilateral humerus and unilateral femur IMN (18; 8 cement and 10 uncemented), followed by unilateral humerus and bilateral femora IMN (9; 2 cement and 7 uncemented).   Figure 1. Flow chart depicting the patient selection process adopted in the current investigation. There were 82 subjects enrolled, accounting for 100 intramedullary nailing procedures. * Reported as mean ± standard deviation (range). § Reported as sample size (%). + As determined by Mirel's criteria [43]. † Although many patients had multiple diffuse lesions, these categories were made based on major lesions that were considered for surgical decision making. Reflecting the comparison of variables between patients according to bone cement usage.
The demographics, except for the following, were similar in both groups ( Table 1). The uncemented group had more impending fractures than the cemented group, who had more completed pathologic fractures. The uncemented group had more diaphyseal lesions, and the cemented group had more proximal lesions. The uncemented group had more patients receiving multiple nails placed in one setting. Within the patients who received multiple nails, 38

Perioperative Protocols and after Care
All patients received coordinated care from a multidisciplinary team, ensuring medical optimization prior to surgery, including a hemoglobin level of at least 10 g/dL. Preoperative embolization of hyper-vascular metastatic tumors was performed in 2 cases (renal cell and hepatocellular carcinomas). Mirels' criteria were mostly used for the prophylactic fixation of impending fractures [43]. The type of IMN (cemented vs. uncemented) predominantly depended on the extent of tumor involvement, bone quality, amount of bone loss, periarticular lesions, extensive and skip lesions, and fixation stability, and it was decided by the senior surgeon (AVM).
The surgical technique is detailed in Supplement S2 [44]. All surgeries were conducted with a minimally invasive technique, except for 7 pathologic fractures that required direct tumor site opening (curettage resection for radioresistant tumors and/or significant bone defects [5 cases], as well as failure to obtain acceptable fracture closed reduction [2 cases]). There were 3 patients who received plate fixation, in addition to intramedullary nailing, for extensive periarticular bone involvement. In addition, one patient in the cemented cohort required intraoperative conversion from a nail to a plate construct due to a proximal nail cutout of the lateral cortex. This patient was excluded from functional score analysis, but they were included in survival and complication analyses. The postoperative rehabilitation protocol was similar for all patients. An arm sling was given for initial postoperative comfort, and immediate range of motion, starting with pendulum exercises, was allowed. Weight bearing depended on bone involvement and fixation stability, but activities of daily living were allowed immediately. After discharge, outpatient visits were scheduled at 2 and 6 weeks postoperatively, every 3 months for the first year, every 6 months for the second year, and once a year afterwards. Radiographs were obtained immediate postoperatively, at 3 months, and at every subsequent visit (Figures 2 and 3). In most patients, chemotherapy and/or radiation (3000-3500 Gy) were started/resumed 10 to 21 days after their index surgery, as deemed appropriate by the multidisciplinary team.
ducted with a minimally invasive technique, except for 7 pathologic fractures quired direct tumor site opening (curettage resection for radioresistant tumors and nificant bone defects [5 cases], as well as failure to obtain acceptable fracture closed tion [2 cases]). There were 3 patients who received plate fixation, in addition to in dullary nailing, for extensive periarticular bone involvement. In addition, one pa the cemented cohort required intraoperative conversion from a nail to a plate co due to a proximal nail cutout of the lateral cortex. This patient was excluded from tional score analysis, but they were included in survival and complication analys postoperative rehabilitation protocol was similar for all patients. An arm sling wa for initial postoperative comfort, and immediate range of motion, starting with pen exercises, was allowed. Weight bearing depended on bone involvement and fixat bility, but activities of daily living were allowed immediately. After discharge, out visits were scheduled at 2 and 6 weeks postoperatively, every 3 months for the fir every 6 months for the second year, and once a year afterwards. Radiographs w tained immediate postoperatively, at 3 months, and at every subsequent visit (Fi and 3). In most patients, chemotherapy and/or radiation (3000-3500 Gy) were sta sumed 10 to 21 days after their index surgery, as deemed appropriate by the mu plinary team. -year-old male with a mixed lytic sclerotic lesion in the proximal meta-diaphyseal region, with a pathologic fracture from a newly diagnosed metastatic prostate cancer. This was treated by a bone biopsy, followed by a cemented IMN with two proximal inter-locking screws and no distal screw, as shown in the (c) AP and (d) lateral humerus radiographs. Cement was used for augmentation, due to poor proximal humerus bone quality, to support the nail and the inter-locking screws.
(a) (b) Figure 2. (a) AP and (b) lateral right humerus radiographs of a 67-year-old male with a mixed lytic sclerotic lesion in the proximal meta-diaphyseal region, with a pathologic fracture from a newly diagnosed metastatic prostate cancer. This was treated by a bone biopsy, followed by a cemented IMN with two proximal inter-locking screws and no distal screw, as shown in the (c) AP and (d) lateral humerus radiographs. Cement was used for augmentation, due to poor proximal humerus bone quality, to support the nail and the inter-locking screws.
Cancers 2023, 15, x FOR PEER REVIEW 6 of (c) (d) Figure 2. (a) AP and (b) lateral right humerus radiographs of a 67-year-old male with a mixed ly sclerotic lesion in the proximal meta-diaphyseal region, with a pathologic fracture from a new diagnosed metastatic prostate cancer. This was treated by a bone biopsy, followed by a cement IMN with two proximal inter-locking screws and no distal screw, as shown in the (c) AP and ( lateral humerus radiographs. Cement was used for augmentation, due to poor proximal humer bone quality, to support the nail and the inter-locking screws. (a) (b) No cement was used, as there was enough proximal and distal bone to su port the nail and the inter-locking screws, and some healing changes were already evident.

Collected Variables
Extracted variables included patient demographics (age, sex, body mass inde [BMI]), primary malignancy diagnosis, fracture type (impending or complete pathologic lesion location, cement use, concomitant procedures (e.g., other long bone IMN in th same or different setting), intraoperative blood loss, and blood transfusion volume (up 24 h postoperatively, 1-unit PRBC = 325 mL). Perioperative complications and the retur to the operating room were recorded. Complications were labeled as "surgical" if the were directly related to the nailing procedure, and they included infection, wound dehi cence, reoperations, and implant-related mechanical complications. Some cement extru sion to soft tissues was common in pathologic fracture cases, but it was only classified a a complication if there were subsequent mechanical issues, neurovascular damage, and/o a need for additional intervention. Medical complications were primarily recorded for th admission and cardiopulmonary, including pulmonary embolism, pneumonia, hypoten sion, myocardial infarction, and respiratory distress. Other recorded complications in cluded gastrointestinal bleeding, Clostridium difficile infection, coagulopathies such a disseminated intravascular coagulation, deep vein thrombosis, and thrombocytopeni sepsis, urinary tract infection, as well as transaminitis with hyperbilirubinemia. Intraope ative estimated blood loss (EBL) was estimated by quantifying the number of laparotom sponges utilized along with the total blood in the suction canister [45]. Functional ou comes were recorded using the Musculoskeletal Tumor Society (MSTS) upper extremi scoring system at the initial presentation and 3 months postoperatively [46]. Due to deat loss to follow-up, or an inability to have a direct follow-up visit, especially during th COVID-19 pandemic, functional score was not available for the majority of patients b yond 3 months, and thus, the MSTS score was reported for 3 months. Oncologic outcome were evaluated with patient survivorship and calculated by tracking patients from su gery until death. Several of our patients were foreign citizens traveling to and from the year-old female with mixed lytic sclerotic lesion in the proximal meta-diaphyseal region, with a pathologic fracture and periosteal reaction in a patient with established metastatic breast cancer. This was treated by an uncemented IMN, with three proximal locking screws and one distal screw, as shown in (c) AP and (d) oblique humerus radiographs. No cement was used, as there was enough proximal and distal bone to support the nail and the inter-locking screws, and some healing changes were already evident.

Collected Variables
Extracted variables included patient demographics (age, sex, body mass index [BMI]), primary malignancy diagnosis, fracture type (impending or complete pathologic), lesion location, cement use, concomitant procedures (e.g., other long bone IMN in the same or different setting), intraoperative blood loss, and blood transfusion volume (up to 24 h postoperatively, 1-unit PRBC = 325 mL). Perioperative complications and the return to the operating room were recorded. Complications were labeled as "surgical" if they were directly related to the nailing procedure, and they included infection, wound dehiscence, reoperations, and implant-related mechanical complications. Some cement extrusion to soft tissues was common in pathologic fracture cases, but it was only classified as a complication if there were subsequent mechanical issues, neurovascular damage, and/or a need for additional intervention. Medical complications were primarily recorded for that admission and cardiopulmonary, including pulmonary embolism, pneumonia, hypotension, myocardial infarction, and respiratory distress. Other recorded complications included gastrointestinal bleeding, Clostridium difficile infection, coagulopathies such as disseminated intravascular coagulation, deep vein thrombosis, and thrombocytopenia, sepsis, urinary tract infection, as well as transaminitis with hyperbilirubinemia. Intraoperative estimated blood loss (EBL) was estimated by quantifying the number of laparotomy sponges utilized along with the total blood in the suction canister [45]. Functional outcomes were recorded using the Musculoskeletal Tumor Society (MSTS) upper extremity scoring system at the initial presentation and 3 months postoperatively [46]. Due to death, loss to follow-up, or an inability to have a direct follow-up visit, especially during the COVID-19 pandemic, functional score was not available for the majority of patients beyond 3 months, and thus, the MSTS score was reported for 3 months. Oncologic outcomes were evaluated with patient survivorship and calculated by tracking patients from surgery until death. Several of our patients were foreign citizens traveling to and from their home country to receive treatment, making data collection difficult for certain variables. Patient death was documented from medical chart review and direct communication with their families and other medical providers [47]. When mortality was not able to be confirmed by available means and no further communication with the patient or their family members was identified in the patient's chart, individuals who missed two consecutive visits without rescheduling were presumed to be lost to follow-up at the date of the last documented visit to the clinic or hospital, and they were subsequently excluded from analysis of survivorship and functional outcomes. By a 3 month postoperative period, 18 (22.0%) patients died and 18 (22.0%) were lost to follow-up. By a 6 month postoperative period, 23 (28.0%) patients died and 25 (30.5%) were lost to follow-up. By a 12 month postoperative period, 27 (32.9%) patients had died and 29 (35.4%) were lost to follow-up, and thus, analysis was done based on available patients at each time point.

Statistical Analyses
Descriptive analyses were performed to summarize patient demographics, operative variables, and patient outcomes. Statistical analyses included two-sided Fisher's exact and Student's t-tests to compare categorical and continuous variables, respectively. Patient survivorship was reported using Kaplan-Meier estimates with the log-rank test. All analyses were performed in R Statistical Software using a p-value of <0.05 as the threshold for statistical significance.
There were two returns to the operating room in the entire study. From the cemented group, 1 patient developed quadriparesis from cervical spine tumor involvement and instability, requiring urgent neurosurgical decompression and fixation 2 days after nailing procedure. Another individual from the cemented group was taken back to the operating room, 2 years after their index surgery, for a symptomatic prominent backed out proximal locking screw that was removed.
In our study population, median and 1 year survival were comparable to those previously reported by other studies (6.0 vs. 6.4-10.6 months and 38.7% vs. 26.7-49%, respectively), although there may be a difference in calculation methodology and follow-ups between studies, impeding direct comparison [2,14,17,21,27,49,50]. In our study, there was no difference in the survival of patients in the cemented vs. uncemented group. With the numbers available, this may signify that survival after metastasis to the humerus has a similar prognosis, irrespective of the extent of the disease. Thus, surgical management mainly stays a part of palliative care, improving pain, function, and quality of life.
Functional outcomes of patients with metastatic humeral disease improved following IMN insertion, as evidenced by a two-fold overall mean MSTS score improvement at 3 month follow-up. Coupled with the relatively low surgical complication rate and adequate functional gain, our study found IMN fixation to be an effective operative modality for our patients. Patients undergoing cemented humeral IMN exhibited significantly lower preoperative MSTS scores (40.2%) than patients with uncemented nail insertion (66.7%) (p = 0.01). While this was likely due to selection bias, both groups were similar with respect to postoperative MSTS scores at 3 months (89.8% and 90.9%), justifying the use of cement in this group. Given that individuals with larger tumor burdens-and resultant poorer bone stock-are more likely to receive cement augmentation (over 70% of pathologic fractures in the current analysis), bone cement may promote functional gain in more advanced humeral diseases. This finding corroborates previous studies that endorse cement for the improvement of fixation stability, pain mitigation, and function restoration, especially in the short term [10,20,[51][52][53].
Perioperative complications occurred in 21.0% of procedures, with medical (14.0%) and surgical (8.0%) complication rates comparable to those reported in the literature (0-26.0% and 0-28.6%, respectively) [Tables 2 and 3, Supplement S1]. The variation in complication rates between studies could be explained by discrepancies in defining adverse events. For instance, while some analyses placed emphasis, mostly, on cardiopulmonary complications, including thromboembolism, others also included gastrointestinal bleeding and postoperative pneumonia as complications [17,18]. Embolic events are notable in IMN fixation because reaming, nail insertion, and cement augmentation may cause an intramedullary pressure surge, potentially increasing the risk of tumor and bone marrow fat liberation, subsequently leading to an intraoperative cardiopulmonary event [28,54]. However, this has, mostly, been studied in the femur and not the humerus. In this study, only 1 (1.0%) case of symptomatic pulmonary embolism was noted, which is consistent with a previously reported incidence of fat (0.3%) and pulmonary (1.3%) emboli following humeral nail insertion [17]. In our patients undergoing cemented IMN, cement was used in a relatively more liquid state without pressurization, which, combined with predominant use of unreamed or minimally reamed smaller diameter (8 mm) nails and the presence of a fracture acting as a natural vent site, could have contributed to the overall low rate of embolic events.
Patients undergoing cemented and uncemented humeral IMN displayed comparable rates of overall (22.6% vs. 19.1%), medical (13.2% vs. 14.9%), and surgical (11.3% vs. 4.3%) complications. Intraoperative surgical complications (4 [4.0%] cases) primarily occurred in individuals treated with cemented IMN, reflecting the higher proportion of pathologic fractures in this patient population, more extensive disease with relatively poorer bone stock, and the procedure's technical complexity. Although our cementing technique increased the probability of extrusion at the fracture sites, many of these at the fracture site had no clinical significance and required no intervention, possibly due to an intact protective soft-tissue sleeve, as most of these fractures are usually low-energy events. When needed, the cement was removed in the same setting by using either the same or a different incision. Adding to the literature, most surgical complications encountered with cemented nails were technical, which were recognized and managed appropriately with no impact on patient outcomes [21,29]. Furthermore, intraoperative blood loss was also higher in subjects with cemented IMN, which usually required taking additional surgical steps, increasing canal reaming, and, possibly, lengthening the operative time as a result, although this extra time was slightly offset by not using distal interlocking screws.
The current study has several limitations. Given its retrospective and complex nature, no true matched control group was included. As such, there were certain confounders of survivorship, functional outcomes, and perioperative complications that could not be accounted and controlled. Several patients underwent multiple long bone nailing procedures, either in a single stage or multiple staged fashion, which can influence overall outcomes. Moreover, there was a selection bias for the usage of cement based on the surgeon's assessment. Long bones requiring prophylactic fixation have relatively better quality, so uncemented intramedullary nails were, more often, inserted for diaphyseal lesions and impending fractures. Diaphyseal lesions also have better proximal and distal bones remaining for fixation and interlocking screw purchasing, thus precluding the need for cement augmentation. Patients with uncemented nails also underwent significantly more than one IMN procedure in multiple long bones in the same setting, creating another bias for complications, as anticipated complexity, surgical time, and blood loss was less. In contrast, complete pathologic fractures often stem from more advanced disease, and as a result, they exhibit inferior bone quality, warranting cement augmentation for better fixation. The lower preoperative MSTS score in subjects who underwent cemented nailing could have mirrored the higher percentage of complete pathologic fractures documented in this patient population. Despite evaluating one of the largest groups of a detailed humeral IMN reported to date (Supplement S1), the sample size included in the current analysis is still relatively small, and it may be limited by insufficient statistical power for the comparison of the two cohorts. Moreover, several patients were lost to follow-up, and they were excluded from the analysis. This study also did not compare operative time directly, as overall surgical time was influenced by several independent factors such as fluoroscopy, perioperative additional anesthesia preparation, and multiple single-stage nailing [42]. Likewise, we did not compare the length of hospital stays and returns to definitive adjuvant therapy in this patient population due to several factors listed above. In addition, most subjects had multiple myeloma as their primary malignancy, due to the disease's high prevalence in our community. Nevertheless, there was no significant difference (p > 0.05) in perioperative complications or survivorship between those with multiple myeloma and those with metastatic disease. Despite similar orthopedic management, metastatic tumors and multiple myeloma may differ regarding disease history and prognosis, potentially skewing the study's findings.
To the best of our knowledge, this study is the largest series of its kind, despite these limitations, and it is the first to compare the survivorship, functional outcomes, and perioperative complications of cemented and uncemented IMN in order to assess their effectiveness as treatments for impending and complete pathologic humeral fractures. Future plans may include a higher-powered analysis, an in-depth analysis of differences between patients with multiple myeloma and metastatic disease, as well as a way to account for other variables and confounders.

Conclusions
Intramedullary nailing, both with and without cement, is a relatively safe and effective therapeutic modality for impending and pathologic humeral fractures, resulting in similarly acceptable clinical outcomes and complication rates. The use of bone cement is often based on several clinical factors and, thus, induces selection bias. Nevertheless, the outcomes were similar between both cemented and uncemented groups, justifying the use of cement in this select group. Most intraoperative surgical complications resulted from technical errors stemming from bone cement use, and they could be minimized with awareness, meticulous attention to surgical technique, and more abundant experience. While controlling for possible selection bias, larger-scale higher-level studies are warranted to validate these results.

Supplementary Materials:
The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/cancers15143601/s1, Supplement S1: Literature review of the last 30 years summarizing complication rates and mean survivorship in patients undergoing intramedullary nailing for metastatic humeral disease. Supplement S2: Surgical technique. Informed Consent Statement: Patient consent was waived due to retrospective study design with deidentified patient information.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical considerations.

Conflicts of Interest:
The authors declare no conflict of interest.