Vertebroplasty and kyphoplasty: national outcomes and trends in utilization from 2005 through 2010

Background context: Vertebral compression fractures secondary to low bone mass are responsible for almost 130,000 inpatient admissions and 133,500 emergency department visits annually, totaling over $5 billion of direct inpatient costs. Although most vertebral compression fractures heal within a few months with conservative therapy, a significant portion fail to improve with conservative treatment and require long-term care, conservative treatment, or both. Fractures that fail conservative therapy are treated with vertebral augmentation procedures (VAPs) such as vertebroplasty (VP) and kyphoplasty (KP). Two large randomized clinical trials published in 2009 questioned the efficacy of VP in treatment of VAPs.

Purpose: This study aimed to investigate trends in utilization of VP and KP between 2005 and 2010 to capture the impact of the 2009 literature on utilization of VAPs. The study also compares patient characteristics and perioperative outcomes between VP and KP to further delineate the risks of each procedure.

Study design: Retrospective analysis of national utilization rates, clinical outcomes, patient demographics, and patient comorbidities using a large national inpatient database.

Patient sample: A total of 63,459 inpatient admissions from 46 states and more than 1,000 different hospitals were included in the analysis.

Outcome measures: Length of stay (LOS), total direct cost, mortality, postoperative complications.

Methods: Data were obtained from the National Inpatient Sample database for the period between 2005 and 2010. National Inpatient Sample is the largest publicly available all payer inpatient database in the United States. Patients undergoing VP and KP were identified via corresponding the International Classification of Diseases, 9th Revision procedure codes. National utilization trends were estimated using weights supplied as part of the National Inpatient Sample dataset. Information on patient comorbidities and demographics was collected. A series of univariate and multivarariate analyses were used to identify statistically significant differences in patient characteristics, clinical outcomes, as well as cost and LOS between patients undergoing VP versus KP.

Results: A total of 307,050 inpatient VAPs were performed in the United States between 2005 and 2010. Of those procedures, 225,259 were KP and 81,790 were VP. Kyphoplasty utilization showed an increasing trend between 2005 and 2007, increasing from 27 to 33 procedures per 100,000 capita older than 40 years. During the same time period, VP utilization remained constant at approximately nine procedures per 100,000 capita older than 40 years. After 2007, utilization of both VP and KP decreased. The most precipitous decrease in VAP utilization occurred in 2009. Patients undergoing VP were on average older (76.7 vs. 77.8, p<.0001), more frequently women (74.48% vs. 73.15%, p=.00083), and black (1.77% vs. 1.55%, p=.004059). Patients undergoing VP had on average more comorbidities then those undergoing KP. Patients undergoing VP had a higher rate of postoperative anemia secondary to acute bleeding and higher rate of venous thromboembolic events. Those undergoing KP had a greater rate of cardiac complications; however, this difference was not statistically significant when taking into account patient age and comorbidity burden. Vertebroplasty was associated with higher mortality (0.93% vs. 0.60%, p<.001), longer LOS (6.78 vs. 5.05 days, p<.0001), and lower total cost ($42,154 vs. $46,101, p<.0001).

Conclusions: Overall, KP was associated with lower complication rates, shorter LOS, and a higher total direct cost compared with VP. Utilization rates showed a significant decrease since 2009 in both VP and KP, suggesting that both procedures were impacted by the two randomized controlled trials published in 2009 that suggested poor efficacy of VP.