OUP user menu

Etiologies of Failed Back Surgery Syndrome

Curtis W. Slipman MD, Carl H. Shin MD, Rajeev K. Patel MD, Zacharia Isaac MD, Chris W. Huston MD, Jason S. Lipetz MD, David A. Lenrow MD, Debra L. Braverman MD, Edward J. Vresilovic Jr. MD, PhD
DOI: http://dx.doi.org/10.1046/j.1526-4637.2002.02033.x 200-214 First published online: 1 September 2002

Abstract

Study Design. Retrospective chart review.

Objective. To report the epidemiologic data of nonsurgical and surgical etiologies of failed back surgery syndrome (FBSS) from two outpatient spine practices.

Summary of Background Data. FBSS has been offered as a diagnosis, but this is an imprecise term encompassing a heterogeneous group of disorders that have in common pain symptoms after lumbar surgery. The current literature primarily diagnoses for the various etiologies of FBSS from a surgical perspective. To our knowledge, there is no study that investigates the myriad of surgical and nonsurgical diagnoses from a nonsurgical perspective.

Methods. Specific inclusion and exclusion criteria were developed for a list of 42 nonsurgical and surgical differential diagnoses of FBSS. The determination of which category, surgical or nonsurgical, each diagnosis was placed into depended upon the categorization of those diagnoses in previously published literature on FBSS. Each of the authors reviewed the definitions, and they came to a unanimous agreement on each diagnosis' inclusion and exclusion criteria. Data extraction was then carried out in each of the two involved institutions by using the key words discectomy, laminectomy, and fusion to identify all the patients who had any combination of low back, buttock, or lower extremity pain after lumbar discectomy surgery. These charts were then individually reviewed to extract epidemiologic data.

Results. A total of 267 charts were reviewed. One hundred and ninety-seven (197) charts had a complete workup. Of these, 11 (5.6%) had an unknown etiology, and 186 had a known diagnosis. Twenty-three (23) various diagnoses were identified. There was approximately an equal distribution between the incidences of nonsurgical and surgical diagnoses; 44.4% had nonsurgical diagnoses and 55.6% had surgical diagnoses. The most common diagnoses identified were spinal stenosis, internal disc disruption syndrome, recurrent/retained disc, and neural fibrosis.

Conclusion. FBSS is a syndrome consisting of a myriad of surgical and nonsurgical etiologies. Approximately one half of FBSS patients have a surgical etiology. Approximately 95% of patients can be provided a specific diagnosis.

  • Failed Back Surgery Syndrome
  • Chronic Low Back Pain
  • Lumbar Surgery
  • Spinal Stenosis
  • Internal Disc Disruption Syndrome
  • Neural Fibrosis
  • Recurrent Disc
  • Retained Disc
  • Discectomy

Introduction

It is estimated that more than five million Americans experience chronic low back pain, with approximately 50% of those afflicted being disabled [1]. Annually, direct medical costs are approximately US$25 billion [2]. The majority of low back pain patients that fail to improve after surgery are classified under the heterogenous disorder most commonly referred to as failed back surgery syndrome (FBSS) [1].

Cases of ruptured intervertebral discs have been reported as early as 1896 [3]. In 1911, Middleton and Treacher [4] of the United Kingdom and J.E. Goldwait [5] of Boston, independently described the entity known as the ruptured intervertebral disc. Then, in 1929, studies by Dandy [6] and Schmorl [7] provided evidence of the possible clinical significance of the ruptured disc, however, its association with sciatica had not yet been elucidated. Subsequently, in 1934, Mixter and Barr [8] reported a ruptured disc causing radicular pain. The following year, Mixter suggested that disc herniations may be symptomatic without obvious radicular involvement [9]. It has been stated that this 1935 article opened the gates for back surgery [10]. Since then, there has been a steady increase in the number of lumbar surgeries performed in the United States.

Between 1979 and 1987, the rates of lumbar laminectomies, discectomies, and arthrodesis increased by 23%, 75%, and 200%, respectively. During the 13-year period following 1987, the rate of operations on the lumbar spine increased an additional 100% to 26 per 100,000 [11]. It is estimated that over 250,000 new laminectomies are performed each year in the United States [11,12]. Each year, approximately 30,000-40,000 of laminectomy patients obtain either no relief of symptomatology or a recurrence of symptoms [13]. However, the frequency of a poor result after lumbar laminectomy surgery is decreased by over 66% compared with the national average when performed in modern comprehensive spine centers [12]. Outcome studies of lumbar disc surgery document a success rate of 49-90%, depending on the evaluation criteria used [14-17]. Results after open lumbar disc operations are poor in 10-30% of patients [12,18,19]. Percutaneous disc surgery revision rates have been reported as high as 65%[20].

FBSS has been offered as a diagnosis; however, this may be inaccurate. In our view, FBSS is an imprecise term. Instead of labeling a specific diagnosis, FBSS encompasses a heterogenous group of disorders that have in common pain symptoms after surgery [21-24]. As such, FBSS should be considered a syndrome with multiple possible explanatory etiologies [2,23,25-29].

The current literature examining the incidence of the various etiologies of FBSS identifies primarily surgical diagnoses [12,13,18,23,27,30]. Since the authors of these investigations are surgeons, such an emphasis is understandable. To our knowledge, there is no study that investigates the myriad of surgical and nonsurgical diagnoses. The purpose of this study is twofold. First, to make an initial attempt at defining nonsurgical etiologies of FBSS. Second, to report the incidence of nonsurgical and surgical etiologies of patients with FBSS following surgery for a herniated lumbar disc.

Methods

A listing of the potential nonsurgical and surgical differential diagnosis of FBSS was constructed. For each diagnosis, specific inclusion and exclusion criteria were developed. These criteria evolved from our own practice standard and incorporated information from published articles from a wide spectrum of medical specialties. Once this initial list of diagnoses and their preliminary definitions was formulated, a round table discussion by the authors of this report was conducted. Each of the authors reviewed the definitions, and they came to a unanimous agreement on each diagnosis' inclusion and exclusion criteria. This dialogue occurred during the 61st Annual meeting of the American Academy of Physical Medicine and Rehabilitation in Washington, DC. Prior to initiating a chart review, a list of the components of the epidemiologic data to be collected was constructed. These components were age, gender, diagnosis, number of previous surgeries, time of most recent surgery, duration of symptom relief after surgery, duration of overall symptoms, visual analog scale (VAS) pain score, location of pain, magnetic resonance imaging (MRI), electrodiagnostic study results, injection results, and provocative discography results.

Data extraction proceeded in a similar manner at each of the two involved institutions. Zip® disks containing consecutive initial patient evaluation dictations stored in Microsoft® Word format were reviewed from a private practice orthopedic group in Arizona (AZ) and an academic spine center (ASC). These evaluations from AZ and ASC spanned 36 and 60 months, respectively. A search was conducted using the key words discectomy, laminectomy, and fusion to identify all the patients who had undergone prior surgery. Patients who had undergone discectomy/laminectomy for a herniated nucleus pulposus were included. Charts were reviewed by an independent reviewer to assess for any combination of low back, buttock, or lower extremity pain after lumbar surgery and to extract epidemiologic data.

Results

The total number of charts reviewed was 267. There were 70 charts with an incomplete workup. The remaining 197 (73.8% of total) charts had a complete workup. Of these, 11 (5.6%) had an unknown etiology despite a complete workup. This resulted in 186 (94.4%) charts with a known diagnosis out of 197 total charts with a complete workup. Of the 186 charts with known diagnoses, six had two diagnoses and three had three diagnoses. This resulted in 186 total charts with 198 total diagnoses. The number of surgical versus nonsurgical diagnoses was 110 versus 88 (55.6% vs 44.4%). A chart of each diagnosis, the raw number of that diagnosis made, and the percentage out of the total number charts with a complete workup are demonstrated in Table 1.

View this table:
Table 1

The absolute number and percentage of each diagnosis as the etiology of FBSS

DiagnosisRaw NumberPercentage*
Surgical
  Stenosis (total)4021.5
    foramina2312.4
    central115.9
    lateral63.2
  Internal disc disruption**4021.5
  Recurrent/retained disc2312.4
  Spondylolisthesis31.6
  Synovial cyst21.1
  Vascular claudication21.1
  Instability10.5
  Pseudomeningocele10.5
Nonsurgical
  Fibrosis (total)2714.5
    epidural158.1
    intraneural126.5
  Degenerative disc disease179.1
  Radiculopathy105.4
  Radicular pain94.8
  Deconditioning73.8
  Facet syndrome52.7
  Battered root syndrome31.6
  Sacroiliac joint syndrome31.6
  Reflex sympathetic dystrophy21.1
  Fibromyalgia10.5
  Discitis10.5
  Arachnoiditis10.5
  Unknown115.6
  • * Percentage is based on number of diagnoses calculated over total number of patients, and not total number of diagnoses. Therefore, the total percentage is greater than 100%. There were 186 patients with 198 diagnoses

  • ** Internal disc disruption syndrome is included in surgical diagnoses although a nonsurgical treatment may be available.

Of the charts that had a complete workup, there were 108 (58.1%) male and 88 (41.9%) female patients. The mean age of these patients was 51.6 years (range 19-84). The mean number of prior surgeries was 1.6, with a range of 1 to 6. The average duration of symptoms prior to initial evaluation was 41.4 months, with a range of 1 to 360 months. The average duration of symptom relief after the most recent surgery was 155 weeks (3.1 years), with a range of 0 to 35 years. Seventy-three (73) of 186 (39.2%) patients experienced less than 20% relief of their back and/or leg symptoms after their most recent surgery. The duration of symptom relief after the most recent surgery for all diagnoses and the four most common diagnoses are demonstrated in Table 2. For spinal stenosis, 57% experienced recurrence of pain within six months of their surgery. In more than one half (54%) of patients diagnosed with internal disc disruption (IDD) syndrome, the symptoms were persistent prior to and following surgery. Those with recurrent/retained disc (Rec/Ret disc), had a bimodal distribution, with 30% experiencing symptoms within six months and 48% experiencing recurrence five or more years after surgery. The majority of the unknown diagnoses (9/11) experienced either persistent symptoms or recurrence within six months of their surgery.

View this table:
Table 2

Duration of symptom relief after the most recent surgery

NNo relief<6 mo6-12 mo1-2 yrs2-5 yrs5-10 yrs>10 yrs
All diagnoses18938.6%21.6%4.8%5.3%9.5%10.6%9.5%
(73)(41)(9)(10)(18)(20)(18)
Stenosis 3930.8%33.3%7.7%5.1%7.7%5.1%10.3%
(12)(13)(3)(2)(3)(2)(4)
IDD syndrome 3953.8%10.3%10.3%5.1%7.7%5.1%7.7%
(21)(4)(4)(2)(3)(2)(3)
Rec/Ret disc 2218.2%13.6%0%4.5%13.6%31.8%18.2%
(4)(3)(1)(3)(7)(4)
Scarring 2556%40%4%0%0%0%0%
(14)(10)(1)
Unknown 1154.5%27.3%0%0%9.1%9.1%0
(6)(3)(1)(1)
  • For patients with multiple diagnoses, duration of symptom relief after the most recent surgery based on their principle diagnosis was used.

Discussion

Numerous published reports have demonstrated that 90-95% of all episodes of low back pain cases are nonsurgical [31-34]. Interestingly, prior epidemiologic studies report that the nonsurgical etiologies of FBSS represent between 0 and 83% of the underlying causes of FBSS. See Table 4 for a comparison of diagnoses attributed to failed back surgery syndrome in prior studies. Burton and Kirkaldy-Willis reported the first epidemiologic study examining the underlying etiologies of FBSS [21]. This combined neurosurgical and orthopedic North American paper reported that the frequency of surgical diagnoses compared with nonsurgical diagnoses was greater than a 2:1 ratio. Burton's 10-year follow-up paper reviewed his prior reported causative factors and their status [12]. He stated that surgical etiologies remained the most frequent cause of FBSS [12]. Subsequently, a neurosurgical study by Long conducted in a population of patients in which 64 out of 78 underwent surgery, reported a 4.5:1 incidence of patients that had a nonsurgical versus surgical etiology of failure. A breakdown of etiologies in the patients who were operated on versus those who were managed nonoperatively was not provided [23,35]. In 1993, Bernard published the first orthopedic study. He indicated that greater than 90% of the causes of FBSS were surgical. In that study, arachnoiditis and degenerative spondylosis were the only nonsurgical etiologies identified [36]. The most recent epidemiologic study of FBSS was presented by Simmons at the North American Spine Societies' annual conference [37]. He reported upon 212 postoperative patients with FBSS. All the underlying diagnoses uncovered were surgical. A review of these papers and presentations reveals that none employed specific inclusion or exclusion criteria for the nonsurgical or surgical diagnoses.

View this table:
Table 4

Comparison of studies on failed back surgery syndrome

B & KWB & KW2BernardLongSimmonsSlipman
Surgical
  Spinal Stenosis, total64722951221.5
  Stenosis, foraminal112.4
  Stenosis, central71429125.9
  Lateral spinal stenosis57583.2
  Internal disc disruption2921.5**
  Severe spondylosis5
  Segmental instability2
  Instability/disc degeneration350.5
  Spondylolisthesis4151.6
  Recurrent or retained HNP12163315912.4
  HNP at a new level7
  Scoliosis17
  Pseudoarthrosis< 52915
  Foreign body< 5
  Surgery performed at the wrong level< 5
  Traumatic meningocele1
  Tarsal tunnel syndrome1
  Fractured hip1
  Compression fracture1
  Synovial cyst1.1
  Vascular claudication1.1
  Pseudomeningocele0.5
Nonsurgical
  Arachnoiditis6*16*11130.5
  Epidural or intraneural fibrosis6*8*1414.6
  Nerve injury during surgery/Battered root< 51.6
  Chronic mechanical pain< 5
  Transitional syndrome< 5
  Unknown< 55.6
  Normal21
  Expected post-op changes21
  Traumatic neuritis6
  Cancer4
  Musculoskeletal abnormality3
  Degenerative spondylosis9
  Mechanical low back pain9.1
  Radiculopathy5.4
  Radicular pain4.8
  Deconditioning3.8
  Facet syndrome2.7
  Sacroiliac joint syndrome1.6
  Complex regional pain syndrome0.5
  Fibromyalgia0.5
  Discitis0.5
  • * It is unclear how the diagnosis of arachnoiditis or epidural fibrosis was defined in this report.

  • ** Internal disc disruption syndrome is included in surgical diagnoses although a nonsurgical treatment may be available.

  • Study by Long included 78 patients, 64 of which had surgery. A breakdown of the etiologies in patients who had surgery versus those that did not was not performed in this study.

While the aforementioned epidemiologic studies have elucidated the various surgical causes of FBSS, there has been no systematic analysis of the nonsurgical etiologies of FBSS. In this report, specific criteria had to be met for a patient to be given a particular diagnosis (Appendix). The value of utilizing specific inclusion and exclusion criteria is manifested by its use in clinical and academic settings. First, it allows the clinician to refer to a comprehensive list from which to offer a patient a diagnosis. Second, it allows for the performance of epidemiologic studies. Third, by using common criteria, outcome studies can be conducted.

We recognized prior to embarking upon this project that the most difficult aspect would be establishing objective definitions for the nonsurgical etiologies. Unlike surgical etiologies, there are no gold standard diagnostic tests with which the nonsurgical etiologies can be compared. Consequently, establishing definitions of nonsurgical etiologies of FBSS is inherently a difficult task. This process is further complicated because different clinicians employ variable criteria for the diagnoses listed in this report. In our opinion, this disparity in definition stems from differences in medical, specialty, and fellowship training; geography; practice setting; patient population; cultural/philosophical views; and individual bias. Although we anticipated this problem, we thought an initial attempt at providing objective defining criteria for the various nonsurgical etiologies to be a useful process.

There are several previously reported diagnoses that were excluded from our listing. There were several reasons for this decision. The authors of this report acknowledge somatoform, conversion reaction, and other psychologic disorders as significant causes of FBSS. These diagnoses were specifically not included in our list of etiologies because the purpose of this study was to identify and report upon the physical causes of FBSS. Similarly, components of the chronic pain syndrome including psychologic, environmental, and socioeconomic issues were omitted from the list of etiologies. It has been suggested that the most frequent etiologies of FBSS include “wrong diagnosis,” “wrong surgeon,” and/or inappropriate patient selection [12,18,21, 36,38]. We do not consider these as diagnoses, but instead, pejorative statements that do not lead to a specific diagnosis or treatment algorithm. Given this view, we did not use these aforementioned “diagnoses.”

The classification of FBSS varies according to the authors [12,13,21,23, 36,39-41]. This difference in categorization of FBSS relates to medical specialty, patient population, and location of practice. Although we have emphasized the differences between reports from various authors, there are similarities, which should be highlighted. This commonality relates to incorporating the time of symptom presentation into the probability analysis of the differential diagnosis. Gill and Frymoyer [2] proposed a classification scheme in which failed back syndrome is divided into four different subcategories based on a temporal probability scale: Immediate failures, early recurrence, midterm failures, and long-term failures predicated upon the duration of symptom relief postoperatively. The common causes of immediate failure include wrong preoperative diagnosis or technical error. The common causes of early recurrence are infection and meningeal cysts. The common causes of midterm failures include recurrent disc prolapse, battered root, arachnoiditis, and paraspinal muscle denervation. The common causes of long-term failures include recurrent stenosis and instability [2]. This model classifying the surgical etiologies of FBSS temporally may be applied to the nonsurgical causes of FBSS.

Our results indicate that 61.2% of all patients experienced back and/or leg pain within six months of their most recent surgery. For spinal stenosis, 57% experienced pain within six months, and 54% of IDD syndrome patients had persistent back pain after surgery. As one would expect, retained/recurrent disc patients had a bimodal distribution in duration of symptom relief after surgery. Thirty (30) percent had pain within six months, and 48% experienced recurrence of their symptoms five or more years later. All patients with neural fibrosis experienced either persistent or recurrent pain within one year, but this was inherently biased, as the inclusion criteria defined fibrosis within this time frame. For those with unknown diagnoses despite complete workup, 9/11 experienced recurrence of pain within six months of their surgery.

Our results demonstrate that FBSS is a syndrome consisting of numerous surgical and nonsurgical etiologies for the population of FBSS patients presenting to a spine practice (Table 3). For that patient population, approximately 95% of patients with FBSS can be provided a specific diagnosis. The ratio of surgical to nonsurgical etiologies is approximately 1:1. This ratio differs substantially from the prior reports of Burton and Kirkaldy-Willis [21], Bernard [36], and Simmons [37]. There are differences in the results of prior published reports compared with our study, which in part relates to available technology, patient population, and the use of discography. In the study by Burton and Kirkaldy-Willis [21], MRI was not available, thereby affecting the accuracy of preoperative diagnoses. The Burton and Kirkaldy-Willis [12] study did not entertain the diagnosis of IDD, which requires confirmation by concordant provocative discography. In contrast, our study and those of Bernard [36] and Simmons included this diagnosis as one of the potential etiologies of FBSS.

View this table:
Table 3

Differential diagnosis based on predominant symptom

Predominantly Back Pain with or without referred pain to the lower limb Predominantly Leg Pain with or without associated low back pain
Back PainRecurrent disc
Infection:Retained disc
  DiscitisLateral stenosis
  OsteomyelitisForaminal stenosis
  Epidural abscessExtraforaminal stenosis
  Soft tissueFar-out stenosis
Facet fractureMeningocele
TumorEpidural hematoma
IDDSeroma
Facet joint painTumor
SI jointSynovial cyst
InstabilityEpidural fibrosis
PseudoarthrosisIntraneural fibrosis
SpondylolysisBattered root syndrome
SpondylolisthesisCentral stenosis
Mechanical low back painArachnoiditis
Buttock PainPiriformis syndrome
Iliac crest donor siteIliotibial band syndrome
Cluneal neuropathyHip pathology
MiscellaneousKnee pathology
Myofascial pain syndromeComplex Regional
Fibromyalgia  Pain Syndrome Type 2
DeconditioningVascular claudication
Unknown

These results carry specific implications for the spine specialist. Approximately one half of FBSS patients will have a surgical etiology. Therefore, the practicing spine specialist must be comfortable with diagnosing a surgical lesion in this patient population. The ability to detect an anatomic lesion by various types of radiographic spine studies, such as flexion/extension roentgenograms, MRI with and without gadolinium, multiplanar reformatted computed tomography (CT), discography, postdiscography CT, and myelography, is required. When a surgical lesion is uncovered that fails to improve with conservative treatment, the spine specialist should not hesitate to make a referral to an orthopedic or neurosurgical spine surgeon.

Generalized statements for the practicing spine surgeon cannot be made based on this study because of differences in patient populations. However, it is reasonable to suggest that for those patients presenting to a spine surgeon for whom there is no surgical lesion present, a specific nonsurgical diagnosis can be provided.

This paper provides an initial framework from which to derive a differential diagnosis for a patient presenting with FBSS. We offer a retrospective analysis of the incidence of these various surgical and nonsurgical etiologies. Further study is required to refine the differential diagnosis listed, and prospective study is needed to confirm the epidemiologic data.

Appendix: Inclusion Criteria

View this table:

Footnotes

  • This study was a multicenter study involving the above authors. They were involved in conception, design, and a round table discussion that was the basis for outlining of the inclusion and exclusion criteria for the 37 diagnoses.

References

View Abstract