1 Introduction

Metastatic bone disease causes a significant financial burden to both patients and hospitals. In 2007, Schulman et al. showed that metastatic bone disease cost $12.6 billion nationally.^1,2 This number has likely grown significantly and will continue to grow with the increasing number of people older than 65 years old in the US population. Thus, identifying patient factors that increase hospital length of stay (LOS) is an important step in decreasing the financial burden of this disease.

Previous research indicates greater numbers of comorbidities lead to increased LOS and overall cost following common orthopedic procedures.^3–5 Specific medical conditions are known to increase LOS following orthopedic surgery including congestive heart failure, diabetes mellitus, COPD, hypertension, morbid obesity, and underweight status.⁶ It is also known that patient demographics such as increasing age and non-white race can negatively affect orthopedic surgical outcomes.^7–9 However, there is a lack of research examining the effect of these variables on patients with metastatic bone disease length of hospital stay and ultimately economic burden following surgery.

While bone metastasis surgery can reduce pain and improve function, a firm understanding of risks is vital to medical decision making. Studies have shown that a shorter hospital LOS leads to better functional outcomes and lower mortality, especially in older patients.^10,11 Institutions also benefit from shorter admissions through decreased resource strain.

The present study seeks to identify factors associated with decreased and increased length of hospital stay, that when addressed, would allow for improved care for patients. Decreasing length of hospital stay would reduce financial burden and improve patient's quality of life. In evaluating factors that increase length of hospital stay, orthopedic surgeons can better counsel patients prior to undergoing procedures, resulting in sounder, informed, shared decision making.

2 Methods

A retrospective chart review was performed at a single institution between 2017 and 2019 in which 93 patients that underwent orthopedic surgery for metastatic cancer were identified. Data was obtained through the electronic patient medical record using CPT codes. Data collected included length of hospital stay; demographic information including age, sex, race, body mass index, smoking status (Table 1); pre-operative lab values including white blood cell count, albumin, creatinine, hemoglobin A1c (Table 2); postoperative complications including infection, deep venous thrombosis, pulmonary embolism, fractures (Table 3); primary cancer type, presence or absence of radiation or chemotherapy; and surgical procedure measures (prophylactic, nonprophylactic) (Table 1) in order to understand which factors affected length of hospital stay. The length of stay was defined as the duration from the day of admission to the day of discharge. The study was approved by the institution's IRB board.

Analysis of covariance (ANCOVA) was used for the statistical analysis, with the length of stay as the outcome. Due to outliers, a log transform was applied to the length of stay data before analysis. The site of the disease and the primary cancer, as well as the interaction between them, were the factors analyzed, with the following covariates: age, sex, chemotherapy, radiation, creatinine range (low, normal, or high), prophylactic procedure, diabetes, hypertension, pulmonary/circulatory disorder, obesity, and WBC range (low, normal, or high). A post-hoc Tukey HSD test was then run to compare all possible pairs of means and determine true significance.

3 Results

The median length of hospital stay was 4.5 days with the duration ranging from 1 to 23 days.

There was no difference in LOS when comparing the primary cancer types (p = 0.053). The interaction between site and primary cancer (p = 0.833) was not significant; therefore, the differences of LOS between sites (hip, humerus, etc) remained, regardless of the type of primary cancer.

A post-hoc Tukey HSD test showed a significantly longer LOS for patients with multiple disease sites than for patients with a single disease site (Δ = 3.036, 95% CI = [1.115, 8.268], p = 0.042). Patients with disease at the femur were shown to have longer LOS than other disease sites (hip, humerus, other) (Δ = 2.320, 95% CI = [1.045, 5.153], p = 0.035). The rest of the pairwise contrasts showed no significant difference.

Prophylactic procedures were associated with a decreased LOS (3.53 days) when compared to non-prophylactic procedures (5.51 days) (p = 0.005). Diabetes and obesity were associated with an increased LOS at 6.33 days for diabetics vs 4.85 days for non-diabetics (p = 0.005) and 6.26 days for obese patients vs 4.47 days for those with a normal BMI (p = 0.025).

4 Discussion

Both patients and institutions benefit from minimizing length of hospital admission following surgery. The median LOS following metastatic bone tumor surgery, in this study, was found to be 4.5 days. Identifying the factors that contribute to increased LOS in metastatic bone tumors is important not only for communicating expectations to patients, but also for mitigation of these factors and proper preparation to avoid additional complications associated with extended hospital stays. Determining the median LOS also allows for better estimation of cost burden for the duration of hospital admission.

The present study reports an increased length of hospital stay in patients undergoing orthopedic procedures for metastatic bone disease with pre-existing diabetes or obesity. This finding supports previous research showing increased surgical risk in patients with multiple comorbidities.¹² Previous research in non-oncologic orthopedics demonstrated that patients with decreased length of hospital stay were found to have lower odds of decline in physical function.^10,11 Those with pre-existing conditions are more susceptible to complex infectious diseases and increased mortality, both of which are compounded by the addition of increased LOS. The results of this study only further emphasize the importance of preventative medicine in decreasing LOS and would suggest better long-term physical outcomes in patients without diabetes or obesity.

LOS is also impacted by the disease site (i.e. femur, hip, humerus, multiple sites, etc.). Those with multiple sites of disease or disease involving the femur were found to have longer hospital stays. In previous research studies, LOS was highly dependent on the specific choice of management for femoral bone metastasis. In a study by Tanaka, internal fixation (IF) resulted in a shorter median time to walking than endoprosthetic replacement (EPR).¹³ A study specifically comparing the outcomes of intramedullary nailing (IMN) and EPR for the treatment of proximal femur lesions showed EPR resulted in an average of an 8-day stay versus 5 days in IMN.¹⁴ It is important to consider individual patient risk factors when determining how to manage femoral lesions and whether a patient will be able to tolerate longer hospital stays associated with specific interventions.

Previous research has shown that prophylactic stabilization at the femur resulted in a lower risk of death compared to patients treated for pathologic femur fracture.¹⁵ Research also shows patients treated for pathologic fractures had higher total cost than patients treated prophylactically.¹⁶ The current study found that patients were found to have a decreased length of stay when undergoing prophylactic procedures (3.53 days vs 5.51 for non-prophylactic). In a study evaluating prophylactic fixation versus post-fracture stabilization in metastatic long bone lesions, prophylactic fixation was associated with significantly lower rates of major medical complications, higher rates of discharge home instead of a care facility, and decreased LOS.¹⁷ The findings of this study reinforce the known benefit of prophylactic stabilization of the femur specifically regarding patient LOS.

There are several limitations to the present study, however. First, the sample size is relatively small compared to similar research published on other orthopedic conditions. Despite the smaller size, however, statistically significant risk factors affecting LOS were identified. Second, the study is retrospective and thus open to selection bias and potentially incomplete information. Third, patients with metastatic cancer have very disparate medical needs outside of their orthopedic procedure which could have added some confounding variables to our analysis.

In conclusion, the present study reports increased LOS in patients undergoing orthopedic procedures for metastatic bone disease who had pre-existing diabetes or obesity, multiple disease sites, disease in the femur, or surgery for pathologic fracture (as opposed to a prophylactic procedure). Understanding the factors affecting LOS in this patient population can optimize preoperative care, improve communication and shared understanding between patient and provider, and decrease financial burden.