Richard A. Deyo, MD, MPH; Sohail K. Mirza, MD, MPH; Brook I. Martin, MPH;
William Kreuter, MPA; David C. Goodman, MD, MS; Jeffrey G. Jarvik, MD, MPH
Department of Family Medicine,
Mail Code FM, Oregon Health and Science University,
3181 SW Sam Jackson Park Rd, Portland, OR 97239, USA.
Trends, Major Medical Complications, and Charges Associated with Surgery
for Lumbar Spinal Stenosis in Older Adults
JAMA. 2010 (Apr 7); 303 (13): 1259–1265 ~ FULL TEXT
CONTEXT: In recent decades, the fastest growth in lumbar surgery occurred in older patients with spinal stenosis. Trials indicate that for selected patients, decompressive surgery offers an advantage over nonoperative treatment, but surgeons often recommend more invasive fusion procedures. Comorbidity is common in older patients, so benefits and risks must be carefully weighed in the choice of surgical procedure.
OBJECTIVE: To examine trends in use of different types of stenosis operations and the association of complications and resource use with surgical complexity.
DESIGN, SETTING, AND PATIENTS: Retrospective cohort analysis of Medicare claims for 2002-2007, focusing on 2007 to assess complications and resource use in US hospitals. Operations for Medicare recipients undergoing surgery for lumbar stenosis (n = 32,152 in the first 11 months of 2007) were grouped into 3 gradations of invasiveness: decompression alone, simple fusion (1 or 2 disk levels, single surgical approach), or complex fusion (more than 2 disk levels or combined anterior and posterior approach).
MAIN OUTCOME MEASURES: Rates of the 3 types of surgery, major complications, postoperative mortality, and resource use.
RESULTS: Overall, surgical rates declined slightly from 2002-2007, but the rate of complex fusion procedures increased 15-fold, from 1.3 to 19.9 per 100,000 beneficiaries. Life-threatening complications increased with increasing surgical invasiveness, from 2.3% among patients having decompression alone to 5.6% among those having complex fusions. After adjustment for age, comorbidity, previous spine surgery, and other features, the odds ratio (OR) of life-threatening complications for complex fusion compared with decompression alone was 2.95 (95% confidence interval [CI], 2.50-3.49). A similar pattern was observed for rehospitalization within 30 days, which occurred for 7.8% of patients undergoing decompression and 13.0% having a complex fusion (adjusted OR, 1.94; 95% CI, 1.74-2.17). Adjusted mean hospital charges for complex fusion procedures were US $80,888 compared with US $23,724 for decompression alone.
CONCLUSIONS: Among Medicare recipients, between 2002 and 2007, the frequency of complex fusion procedures for spinal stenosis increased while the frequency of decompression surgery and simple fusions decreased. In 2007, compared with decompression, simple fusion and complex fusion were associated with increased risk of major complications, 30-day mortality, and resource use.
From the Full-Text Article:
Rates of surgery for lumbar stenosis declined slightly from 2002-2007, but use of more complex procedures increased substantially. More complex procedures were associated with greater complications, mortality, hospital charges, and other measures of health care use, even after adjustment for patient demographic and clinical characteristics. Age was less predictive than comorbidity or type of surgical procedure.
It is unclear why more complex operations are increasing. It seems implausible that the number of patients with the most complex spinal pathology increased 15-fold in just 6 years. The introduction and marketing of new surgical devices and the influence of key opinion leaders may stimulate more invasive surgery, even in the absence of new indications.  Surgeons may believe more aggressive intervention produces better outcomes. Improvements in surgical technique, anesthetic technique, and supportive care may make more invasive surgery feasible when risks formerly would have been prohibitive. Financial incentives to hospitals and surgeons for more complex procedures may play a role as may desires of surgeons to be local innovators.
Geographic variations in spine surgery rates are among the largest observed for surgical procedures, and variations in use of fusion surgery exceed those for decompression alone. [1, 25] Such variations persist despite extensive research in this area, in part because of the difficulty of conducting randomized surgical trials. They suggest a poor consensus on indications for surgery or the choice of particular procedures. Studies among spine surgeons indicate substantial variability in decisions to operate, perform a fusion, or use surgical implants. [2-3] Our study shows clinically important consequences of these choices.
Evidence for greater efficacy of more complex procedures for lumbar stenosis is lacking.  For patients who also have spondylolisthesis or scoliosis, spinal fusion may improve outcomes over decompression alone. [8, 27] However, trials establishing an advantage of surgery over nonsurgical care for stenosis alone focused overwhelmingly on decompression without fusion. [5-6] Some trials for lumbar stenosis suggest equivalent efficacy for decompression alone vs decompression and fusion in the absence of spondylolisthesis. 
It is not surprising that fusion procedures are associated with more complications than decompression alone. Compared with decompression, spine fusion requires more extensive dissection, decortication of bone, and longer operative time, and often involves placement of implants. This study confirms previous findings that fusion is associated with greater complications and postoperative mortality than decompression alone. [11, 28]
For other indications, randomized trials suggest that fusion by a single approach with bone grafting alone, fusion with implants, and combined anterior and posterior fusion with implants have similar efficacy for improving pain and function. [4, 12] For patients with stenosis and degenerative spondylolisthesis, fusions with and without implants have similar clinical outcomes. [8, 29-30] However, more complex procedures are associated with more complications. [12, 29, 31] Complications also increase with more operated levels,  and with revision surgery. [7-10,33-34] Our data indicate that these patterns hold true for older patients with spinal stenosis.
Patient demographic and clinical characteristics are generally not matters of choice, but surgeons and patients control the choice of surgical procedure. In the absence of compelling data showing better pain relief or function with more complex surgery, our results may suggest using the least invasive procedure that accomplishes clinical goals. This contrasts with a competing theory that surgeons should correct every anatomic abnormality, hoping to avoid future symptoms. The theory behind this “prophylactic” approach remains unproven, and the risk of greater complications from more extensive surgery must be weighed against potential benefits. Thus, it may be prudent to consider whether decompression alone is sufficient; whether stabilizing structures such as facet joints or interspinous ligaments can be preserved; and if a fusion is planned, how much instrumentation and graft material supplementation is needed.
Our study has the advantage of including all Medicare patients having surgery for spinal stenosis, and not selected patients, centers, or surgeons. It includes nearly complete data on repeat hospitalizations and mortality. However, there are also limitations. Diagnoses and procedures may be miscoded, even though the data are used for billing and subject to audit. Furthermore, spine operations appear to be generally coded accurately.  Surgeons use varying definitions of spine instability, and ICD-9 diagnosis codes may not represent this concept well. The level of detail in ICD-9 spine surgery codes is limited, and information on use of implants is incomplete. Complications may not be consistently recorded, but surgical complications are more reliably coded in large databases than complications from medical therapy.  Furthermore, the complications we examined are more consistently coded than minor complications.  The specificity of claims data for complications is high (unlikely to be coded without a real complication), although sensitivity may be lower (some complications not coded). Thus, we are more likely to underestimate than overestimate complication rates.
Surgeons tailor operations to the nature, extent, and location of an individual’s pathology, but claims data do not indicate severity or extent of anatomic changes, patient symptoms, or functional status. Nevertheless, studies report substantial variability in surgical decision making, even for similar patients. [1-3] Furthermore, accounting for coexisting spondylolisthesis or scoliosis did not alter our results.
Another limitation is that we have presented hospital charges rather than actual resource costs or reimbursements, which typically are substantially less than charges. The relationship between costs and charges is complex and varies by hospital and by type of service.
Among Medicare recipients, between 2002 and 2007, the frequency of complex fusion procedures for spinal stenosis increased, while the frequency of decompression surgery and simple fusions decreased. In 2007, compared with decompression, simple fusion and complex fusion were associated with increased risk of major complications, 30-day mortality, and resource use.