Incidence and prevalence of surgery at segments adjacent to a previous posterior lumbar arthrodesis

Abstract

Background context

Adjacent segment disease (ASD) after lumbar spinal fusion has been an important reason behind the development of nonfusion stabilization technology. However, the incidence, prevalence, and factors contributing to adjacent segment degeneration in the lumbar spine remain unclear. A range of prevalence rates for ASD have been reported in the lumbar spinal literature, but the annual incidence has not been widely studied in this region. Conflicting reports exist regarding risk factors, especially fusion length.

Purpose

To determine the annual incidence and prevalence of further surgery for adjacent segment disease (SxASD) after posterior lumbar arthrodesis and examine possible risk factors.

Study design

Retrospective cohort study.

Patient sample

Nine hundred twelve patients who underwent 1,000 consecutive posterior lumbar interbody fusion procedures, with mean follow-up duration of 63 months (range, 5 months–16 years).

Outcome measures

Further surgery for ASD or surgery-free survival.

Methods

A postal and telephone survey. Follow-up rate: 91% of patients. The annual incidence and prevalence of ASD requiring further surgery were determined using Kaplan–Meier survivorship analysis. Cox proportional-hazards (Cox) regression was used for multivariate analysis of possible risk factors. Significance was set at p<.05.

Results

Further surgery for ASD occurred following 130 of 1,000 or 13% of procedures at a mean time of 43 months (range, 2.3–162 months). The mean annual incidence of SxASD over the first 10 years, in all patients, was 2.5% (95% confidence interval [95% CI], 1.9–3.1) with prevalences of 13.6% and 22.2% at 5 and 10 years, respectively. Cox regression modeling found that the number of levels fused (p≤.0003), age of the patient, fusing to L5, and performing an additional laminectomy adjacent to a fusion all independently affect the risk of SxASD. The mean annual incidence figures in the first 10 years after a lumbar fusion were 1.7% (95% CI, 1.3–2.2) after fusion at single levels, 3.6% (2.1–5.2) after two levels, and 5.0% (3.3–6.7) after three and four levels. The 5- and 10-year prevalences were 9% and 16%, 17% and 31%, and 29% and 40% after single-, two-, and three-/four-level fusions, respectively. The risk of SxASD in patients younger than 45 years was one-quarter (95% CI, 10–64) the risk of patients older than 60 years (p=.003). A laminectomy adjacent to a fusion increases the relative risk by 2.4 times (95% CI, 1.1–5.2; p=.03). Stopping a fusion at L5 is associated with a 1.7-fold increased risk (95% CI, 1.2–2.4; p=.007) of SxASD compared with a fusion to S1, for fusions of the same length.

Conclusion

The overall annual incidence and predicted 10-year prevalence of further surgery for ASD after lumbar arthrodesis were 2.5% and 22.2%, respectively. These rates varied widely depending on the identified risk factors. Although young patients who underwent single-level fusions were at low risk, patients who underwent fusion of three or four levels had a threefold increased risk of further surgery, compared with single-level fusions (p<.0001), and a predicted 10-year prevalence of 40%.

Introduction

Evidence & Methods

Lumbar fusion is commonly performed for degenerative conditions and isthmic spondylolysis. A relatively high reoperation rate has been demonstrated with extension of fusion to adjacent levels in many of these surgeries. Risk factors for progressive disease at adjacent segments have not been well delineated.

In this retrospective study of patients undergoing posterior lumbar interbody fusion, the authors found a mean 2.5% annual incidence of surgery for adjacent-level degenerative processes. However, older patients having longer initial fusions, fusions to L5, and adjacent-level laminectomies appeared to be associated with an increased rate of additional surgery.

This study emphasizes that not all fusions, even if one technique is considered (PLIF), have similar downstream risk for adjacent-level disease. The information is helpful for informed consent. However, some considerations remain. The authors' surgical technique (PLIF) may have impacted their findings. Using repeat surgery as a surrogate indicator of progressive disease is problematic and may underestimate disease progression in older patients (ie, it is possible fewer patients will, or be able to, have additional surgery) or overestimate the disease progression in patients with ill-defined pain syndromes (ie, discogenic pain) who tend to aggressively pursue surgical care.

—The Editors

Thirty years ago, Ehni asserted that, “fusion generates a conflict between immediate benefit and late consequences” [1]. There remains a widely held belief that creation of rigid sections in the lumbar spine will lead to excessive stress on and premature degeneration of the motion segments adjacent to an arthrodesis [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23]. This belief is supported by in vitro evidence of increased stresses and intradiscal pressures at segments adjacent to a lumbar spinal fusion [24], [25], [26], [27], [28], [29]. If true, it is important that clinicians, their patients, and health economists have an accurate knowledge of the incidence and risk factors for the development of adjacent segment disease (ASD) after lumbar arthrodesis. Such knowledge will enable a better understanding of the role and place for nonfusion stabilization technology [21], [30], [31], [32].

A number of clinical studies have examined the prevalence of symptomatic ASD at various time points after lumbar fusion [2], [3], [4], [5], [15], [17], [22], [23], [33], [34], [35]. Harrop et al. [36] recently reported the results of a systematic review of the published rates of radiographic adjacent segment degeneration and symptomatic ASD after lumbar spinal fusion. They emphasized the difference between these two conditions and noted a range in the reported prevalence rates of ASD, between 0% and 36%, over various time intervals.

In 1999, Hilibrand et al. [37] reported the results of their detailed study of both the incidence and prevalence of ASD in 374 patients after anterior cervical fusion. They defined “annual incidence” as the percentage of patients in whom new disease developed during the course of a given year of follow-up, having been disease free at the start of that year, and found this to be 2.9%. They defined “prevalence” as the percentage of all patients in whom symptomatic ASD developed within a given period of follow-up. With the exception of a limited investigation by Ghiselli et al. [3], a similar study to that of Hilibrand et al. does not appear to have been carried out in the lumbar spine.

The aim of the present study was to determine annual incidence and prevalence figures for the lumbar spine in a large consecutive patient cohort. The substantial cohort size and uniform surgical technique aided subgroup analysis and investigation of potential risk factors for surgery for adjacent segment disease (SxASD). Similar Kaplan–Meier (K-M) [38] statistical methodology to Hilibrand et al. [37] was used to determine annual incidence and prevalence. Potential surgical and demographic risk factors for SxASD were explored using Cox proportional-hazards (Cox) regression modeling [39]. To reduce potential inaccuracy in retrospective diagnosis of clinically significant ASD, the authors chose to examine rates of surgery for ASD using the end point of further surgical intervention rather than the presence of symptoms or radiographic signs of degeneration.