What will anesthesiologists be anesthetizing? Trends in neurosurgical procedure usage

To anticipate future changes to the practice of neuroanesthesia, we examined the nationwide trends in frequently performed operative neurosurgery.

We used the Nationwide Inpatient Sample, a random sampling of approximately 20% of United States hospitals from 1993 to 2007. Eight categories of operative neurosurgery were developed, based on procedure codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). We tabulated total volume, mean length of stay, and inhospital mortality and calculated growth rates over the entire period, the most recent 5-year period, and the most recent 2-year period. We examined annual volumes from 1993 to 2007 for each neurosurgical category for the most common ICD-9-CM procedure code within each category.

Intracranial endovascular procedures had the highest compound annual growth rate over the entire study period (32%), the most recent 4 years (29%), and the most recent 2 years (12%). Craniotomy for vascular surgery decreased over the entire period (-4.2%). Spinal fusion had the largest absolute increase over the study period (from 54,000 in 1993 to 350,000 in 2007). All categories except craniotomy for vascular surgery had decreased length of stay across the study period, with compound annual growth rates of -1.2% (ventricular and thecal shunt) to -6.6% (deep brain stimulation). Intracranial endovascular procedures had a much higher growth rate of length of stay than vascular surgery over the most recent 2-year (14% vs 1.0%) and 5-year periods (5.6% vs 1.5%).

The highest volume trends in operative neurosurgery are for spinal fusion (increasing at a rate of approximately 12,000 procedures/y), craniotomy for tumors and other purposes (increasing at a rate of approximately 2,700 procedures/y), and a decrease in shunts (decreasing at a rate of approximately 3,000 per year). The data suggest that intracranial endovascular treatment remains relatively rare, but it is growing exponentially, and lengths of stay are increasing, whereas inhospital deaths are decreasing. The conclusions of this study are limited by the imprecise nature of the ICD-9-CM procedure codes, the categorization scheme we used, and by the sampling methods of the National Inpatient Sample.