Trends in Inpatient Vertebroplasty and Kyphoplasty Volume in the United States, 2005-2011: Assessing the Impact of Randomized Controlled Trials.

STUDY DESIGN

Retrospective analysis of the Nationwide Inpatient Sample, 2005-2011.

OBJECTIVE

To identify trends in procedural volume and rates in the time period surrounding publication of randomized controlled trials (RCTs) that examined the utility of vertebroplasty and kyphoplasty.

SUMMARY OF BACKGROUND DATA

Vertebroplasty and kyphoplasty are frequently performed for vertebral compression fractures. Several RCTs have been published with conflicting outcomes regarding pain and quality of life compared with nonsurgical management and sham procedures. Four RCTs with discordant results were published in 2009.

MATERIALS AND METHODS

The Nationwide Inpatient Sample provided longitudinal, retrospective data on United States' inpatients between 2005 and 2011. Inclusion was determined by a principal or secondary International Classification of Diseases, Ninth Revision, Clinical Modification code of 81.65 (percutaneous vertebroplasty) or 81.66 (percutaneous vertebral augmentation; "kyphoplasty"). No diagnoses were excluded. Years were stratified as "pre" (2005-2008) and "post" (2010-2011) in relation to the 4 RCTs published in 2009. Patient, hospital, and admission characteristics were compared using Pearson χ test.

RESULTS

The estimated annual inpatient procedures performed decreased from 54,833 to 39,832 in the pre and post periods, respectively. The procedural rate for fractures decreased from 20.1% to 14.7% (P<0.0001). Patient and hospital demographics did not change considerably between the time periods. In the post period, weekend admissions increased (34.2% vs. 12.4%, P<0.0001), elective admissions decreased (21.4% vs. 40.0%, P<0.0001), routine discharge decreased (33.0% vs. 52.1%, P<0.0001), and encounters with ≥3 Elixhauser comorbidities increased (54.5% vs. 39.1%, P<0.0001).

CONCLUSIONS

The absolute rate of inpatient vertebroplasty and kyphoplasty procedures for fractures decreased 5% in the period (2010-2011) following the publication of 4 RCTs in 2009. The proportion of elective admissions and routine discharges decreased, possibly indicating a population with greater disease severity. Although our analysis cannot demonstrate a cause-and-effect relationship, the decreased inpatient volume and procedural rates surrounding the publication of sentinel negative RCTs is clearly observed.