Abstract

ABBREVIATIONS

• ACS-NSQIP

  American College of Surgeons National Surgical Quality Improvement Program

• ASA

  American Society of Anesthesiologists

• CI

  confidence interval

• CPT

  current procedural terminology

• ICD

  International Classification of Diseases

• OR

  odds ratios

Lumbar disc herniation is a condition in which the annulus fibrosus of the vertebral disc tears, enabling the disc nucleus to extrude through the fibers and potentially compress the nerves around the disc. This condition typically presents with unilateral radicular pain but can cause more severe symptoms, such as severe bilateral sciatica and motor weakness or the disc can even compress the cauda equina causing weakness, saddle anesthesia, and incontinence.^1,2 In 2012, there were 111 665 total discharges for lumbar disc displacement surgery, costing the United States over 6 billion in aggregate charges. It is likely that the increase of lumbar disc surgeries in the elderly will increase as the population ages.

Both medical and surgical management are often viable options. The primary rationale for lumbar decompression surgery is to relieve nerve root or cauda equinal irritation or compression due to the herniated disc material. While motor manifestations and cauda equina syndrome are far more likely to be managed surgically, there is more debate regarding the role of surgical treatment options among patients with only radicular pain. Clinical studies have shown health benefits in back and leg pain for those undergoing surgical intervention.^3,4 Tosteson et al⁵ determined that surgery was more costly, yet led to greater quality-adjusted life years, vs nonoperative treatment for patients in their forties with a lumbar disc herniated. However, there is a debate regarding how to manage a herniated lumbar disc in the elderly, particularly when pain is the only presenting symptom.

Clinical reports on the outcome of lumbar disc surgery show good results in young or middle-aged patients under 65 yr of age,^6,7 but few studies report on the safety profiles of lumbar disc herniation surgery amongst elderly patients greater than 80 yr old.⁸ These patients likely represent a special cohort with a separate risk profile that remains to be fully characterized. In the current study, a large national surgical database, the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), was utilized to assess and compare the morbidity in patients greater than 80 yr of age (80+) to those of middle age (45-65) undergoing single-level decompressions for lumbar disc herniation for radicular pain.

METHODS

Data Source

A prospective, validated multicenter surgical database from the ACS-NSQIP was used to select patients. Institutional review board approval was not needed for this study due to the de-identified nature of patient information in this online database. About 500 community and academic hospitals in the United States provide preoperative and 30-d postoperative data on consenting, randomly-assigned patients. This includes follow-up postdischarge. Patients older than 18, who underwent major surgical cases between 2006 and 2013, are included.⁹ All criteria for inclusion and exclusion can be found at the ACS-NSQIP website (http://site.acsnsqip.org/about/). Data are collected through written and oral communication with the selected patients and medical chart review. Variables include comorbidities, demographics, current procedural terminology (CPT) code for the surgical procedure, intraoperative parameters, International Classification of Diseases (ICD-9) code for diagnosis, and postoperative complications. These variables are captured with approximately 95% success and less than 1.6% disagreement by the NSQIP database.¹⁰ Each institution participating retains a clinical reviewer in surgery trained for data collection in a standardized manner; routine audits of participating hospitals are also conducted to ensure proper data collection and compliance with Health Insurance Portability and Accountability Act protocols.¹¹

Patient Selection

Patients were first selected for 1-level lumbar decompression based on the procedure with the CPT code: 63005 (laminectomy with exploration and/or decompression of spinal cord and/or cauda equina, without facetectomy, foraminotomy or discectomy [eg, spinal stenosis]; lumbar, except for spondylolisthesis), 63030 (laminotomy [hemilaminectomy], with decompression of nerve root[s], including partial facetectomy, foraminotomy and/or excision of herniated intervertebral disc; lumbar), and 63047 (laminectomy, facetectomy, and foraminotomy [unilateral or bilateral with decompression of spinal cord, cauda equina and/or nerve root[s], [eg, spinal or lateral recess stenosis]], single vertebral segment; lumbar). Cases included patients undergoing decompression for lumbar stenosis from ligamentous hypertrophy as well as radiculopathy from lateral recess/neuroforaminal stenosis. Decompression procedures for spondylolisthesis (CPT 63030) or re-exploration (CPT 63042) were not included.

Then, only patients with the ICD-9 code 722.10, “displacement of lumbar intervertebral disc without myelopathy” were retained for analysis. Other ICD-9 codes corresponding for unspecified intervertebral disc disorders or those with myelopathy were excluded, as this analysis sought to focus on patients who were most likely to be undergoing elective cases for disc herniation. Thus, this cohort was comprised exclusively of patients who received a single-level lumbar decompression procedure for disc displacement without myelopathy.

Preoperative and Postoperative Variables

The ACS-NSQIP user guide defined preoperative comorbidity and postoperative outcome variables.¹¹ Postoperative complications include 30-d mortality, myocardial infarction, return to the operating room, sepsis, deep vein thrombosis, transfusions, cardiac arrest necessitating cardiopulmonary resuscitation, coma greater than 24 h, urinary tract infection, acute renal failure, use of ventilator greater than 24 h, pulmonary embolism, pneumonia, wound dehiscence, and postoperative infection. Only variables that had complete information for all cases were included. Neurology-specific variables such as meningitis were not available in this general surgery database. For simplicity and applicability, a few variables were combined or modified: postoperative surgical site infection was defined by superficial infections—skin and subcutaneous tissue—and deep infections which were beneath the subcutaneous tissue; and diabetes was classified based on oral medication or insulin use. Superficial thrombophlebitis was not considered to be a deep vein thrombosis. Bleeding disorder was defined as any condition that placed the patient at risk for excessive bleeding (including preoperative warfarin or heparin) excluding persistent aspirin therapy.

Statistical Analysis with Age

This study's primary objective¹² was to assess the effects on patients greater than 80 yr old, the 80+ group, vs middle-aged individuals on early postoperative complications after single-level lumbar decompressions for lumbar disc herniation. The middle-aged group included patients between 46 and 65 yr, as this was approximately the mean age of patients receiving the lumbar disc surgery. Of note, nonagenarian patients’ ages are recorded by the NSQIP database as 90 yr exactly to protect patients’ privacy. A control patient group with ages between 45 and 65 yr was included for comparative purposes. Descriptive variables and frequencies were used to provide data about the entire cohort. Chi-square analyses or Fischer's exact tests, when appropriate, were used to compare categorical variables between groups, including preoperative comorbidities and postoperative complications. Binary logistic regression analysis was performed for mortality adjusting for American Society of Anesthesiologists (ASA) class for the 2 cohorts. Adjusted odds ratios (OR) and confidence interval (CI) were reported for the regression. Predictability of our regression models was assessed via the c statistic and goodness of fit was further assessed with the Hosmer–Lemeshow test. Analysis was performed using SPSS 22.0 statistical software (IBM Corp, Armonk, New York). Significance level was set at P < .05.

RESULTS

Preoperative Comorbidities

The NSQIP patient database from 2006 to 2013 included n = 9451 patients who received a single-level lumbar decompression procedure for disc displacement without myelopathy. In our study, 485 patients were 80 yr or older, 23 patients were 90 yr or older, and 8966 patients were between 46 and 65 yr. Demographic statistics are available in Table 1. Overall, the preoperative comorbidities and preoperative characteristics were greatly different between the middle-aged and the 80+ cohorts, with many more preoperative comorbidities in the older patients (Table 2).

Postoperative Comorbidities

The middle-aged and 80+ cohort had no difference among the majority of postoperative complications (Table 3). Of note, the 80+ cohort had significantly greater morbidity with regards to incidence of pulmonary embolism (P = .037, OR: 3.54, CI: 1.21-10.36), intra/postoperative transfusion requirement (P = .01, OR: 2.63, CI: 1.3-5.32), urinary tract infection (P = .011, OR: 3.610, CI: 1.50-8.70), and mortality (P = .046, OR: 7.42, CI: 1.44-38.35).

Logistic Regression

Binary logistic regression demonstrated that when controlling for comorbidities through ASA that age alone is not a predictor for mortality (Table 4). No increased risk of mortality was seen with ASA class I (P = .001, OR: 0.000), and only a slight increase in mortality was seen with ASA class 2 (P = .001, OR: 3.502). Increased mortality was seen with ASA class 3 (P = .001, OR: 11.132) and ASA class 4 (P = .001, OR: 21.990).

DISCUSSION

In this large sample (n = 9451) of patients, who received single-level lumbar decompression procedure for disc displacement without myelopathy, the 80+ age cohort showed significantly greater morbidity with pulmonary embolism, intra/postoperative transfusion requirement, urinary tract infection, and 30-d mortality when compared to the middle-aged cohort. However, the degree of increased mortality is very small: 0.4% in 80+ age cohort vs 0.1% in middle-aged cohort (Table 3). This shows that surgery may be more risky in the elderly population, but overall risk is low.

Pulmonary Embolism/Intra/Postoperative Transfusion Requirement

One interesting finding was a greater incidence of pulmonary embolism within the 80+ cohort. This is consistent with other clinical studies that showed the risk of pulmonary embolism increases in general with age.^12,13 This may be due to an increased incidence of venous thromboembolism, with advancing age, with pulmonary embolism, representing an increasing proportion of these total venous thromboembolism events.¹⁴ Another possible reason could be increased venous stasis secondary to delayed mobilization.

Interestingly, the 80+ cohort required greater intra/ postoperative transfusions, although the overall increase was low (1.9% in the 80+ cohort and 0.7% in the middle-aged cohort). Transfusion is generally indicated at hemoglobin less than 6 g/dL, though other patient-related factors such as age, comorbidities, and risk of ischemia may indicate a transfusion at higher hemoglobin levels. Elderly patients generally have more comorbidities than younger patients, which may increase their susceptibility to the negative effects of anemia and increase their need for transfusion.¹⁵ Blood transfusions have also been linked to higher short- and long-term mortality, increased hospital length of stay, and an increased incidence of morbidities such as atrial fibrillation, renal failure, and systemic inflammatory response syndrome in surgical and nonsurgical patients.^16,-29

Urinary Tract Infections

The 80+ cohort showed significantly greater urinary tract infections (P = .011, OR: 3.61), although the difference between the 2 cohorts was only 0.9% (Table 3). Our finding is supported by Ruben et al,³⁰ who demonstrated that the elderly are at an increased risk for urinary tract infections. This may be due to a waning immune system or delayed recovery time and mobilization necessitating longer periods of catheterization.

Mortality

Elderly patients undergoing lumbar disc surgery were associated with greater 30-d mortality as compared to that of the middle-aged patients. However, the overall mortality rate was low with the 80+ cohort at 0.4% (Table 3). Larger sample sizes may be needed to achieve a higher power. Few other studies have studied the influence of age on the outcome following lumbar disc surgery. Weber et al³¹ reported that older age correlated with less satisfaction in pain control in a randomized comparison of surgical and nonsurgical treatment. Weber's study,³¹ however, did not include patients older than 60 yr of age. In contrast, Fujii et al⁸ showed that elderly patients who underwent lumbar discectomy had similar clinical outcomes, assessed via the Japanese Orthopedic Association score, to younger patients. This study classified the ‘elderly group’ as patients older than 65 yr (n = 12) and had a total n of 37. No patients were older than 75 yr of age.⁸ Our study differed by studying a larger cohort and defined elderly patients as 80 yr or older.

Nie et al³² found lumbar disc surgery to be a safe and acceptable treatment without significant morbidity for octogenarians. The study looked at 90 patients and concluded that there were no significant differences between elderly patients (n = 45, 80 yr or older) and middle-aged patients (n = 45, 40-60 yr) in surgery complication, operating time, and estimated blood loss in 1-level lumbar decompression. Although elderly patients were found to have a significantly increased number of preoperative comorbid conditions (ie, joint problems and pulmonary disease), the only significant difference between the 2 groups was in length of surgery.³² Although our study suggested greater morbidity and mortality in a larger cohort of elderly patients, the overall risk was low. Nie et al's study³² focused on a single spine center while ours looked at a validated, prospective multicenter surgical database.

Single-level decompression for lumbar disc herniation is an elective, pain-related disc disease. Binary logistic regression (Table 4) showed that age alone is not a predictor for mortality once controlling for comorbidities. Patients in ASA class 1 and 2 have little to no increased morbidity and mortality with surgery, and these patients may be best managed surgically. Patients in ASA classes 3 and 4 have some increased mortality; careful selection of these patients towards surgical management is warranted.

Limitations

This study had several limitations that merit attention. The study was retrospective in nature as it utilizes the NSQIP patient database. As detailed in other studies,³³ participation in NSQIP is voluntary, and the sites that participate are not necessarily representative of all American hospitals. Data are not specific to any particular surgeon or site, and outcome data are limited to 30 d of follow-up. Variables extracted into the NSQIP database are selected for their broad applicability to all surgical procedures and do not capture all complications specific to lumbar disc herniation. Given these limitations, it is important to note that the NSQIP program provides 30-d outcomes for a myriad of surgical procedures from a large number of hospitals. This data can then be used to compare outcomes across a variety of patient populations.

Our findings cannot be generalized to elderly individuals who are atypically healthy, as roughly a third of our patients had comorbidities. Finally, the 80+ cohort had an increased risk of mortality, but due to database limitations as discussed in other studies,³⁴ the precise cause of death could not be determined. We believe that this is not the result of age alone, but a combination of all the comorbidities that accompany it.

CONCLUSION

In this large sample of patients who received single-level lumbar decompression procedure for disc displacement without myelopathy, elderly patients had a statistically significant increase in morbidity and mortality, but the overall risks of complications remains low. Physicians should preoperatively counsel patients in ASA class 3 and 4 accordingly about the significantly increased but overall low risks for lumbar disc surgery.

Disclosures

The ACS-NSQIP and the hospitals participating in the ACS-NSQIP are the source of the data used herein; they have not verified and are not responsible for the statistical validity of the data analysis or the conclusions derived by the authors. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

REFERENCES

Neurosurgery Speaks! Audio abstracts available for this article at www.neurosurgery-online.com.

COMMENT

Despite the rising rates of lumbar spine surgery in the elderly, the evidence available to support clinical decision making is limited by the frequently restrictive eligibility criteria of prospective studies.¹ The authors' broadly generalizable findings may therefore guide clinicians providing care to this growing patient population. Their findings, that the risk of complications following single-level lumbar decompression remains low in patients 80 years of age and older, help to define the role of spine surgery among the elderly.

Daniel Yavin