Introduction

Low back pain is the most common cause of disability for adults, and its lifetime prevalence was estimated from 59 to 84% [1]. Degenerative lumbar spine disease (lumbar spondylolisthesis, stenosis, disc herniation or disc diseases) is the most common aetiology of low back pain and can have profound effects on functionality and quality of patient life [2,3,4]. When conservative treatment fails, symptoms of lumbar spine diseases are relieved by appropriate lumbar decompression procedures, such as laminectomy and discectomy [5, 6]. However, decompression may compromise the structure of lumbar segments and lead to further degeneration, abnormal motion, or deformity [7]. Lumbar fusion, which intends to relieve back pain attributed to movement at degenerated joints and increase the foramen space, is considered as a stabilizing treatment that may reduce the need for additional surgery [8]. Lumbar fusion surgeries have grown dramatically in the past decades [9, 10]. The number reached 245,000 in 2011 in US alone, while lumbar discectomies occurred in approximately 197,000 inpatients in 2011 [11].

However, it is controversial on the clinical advantages of the decompression alone and decompression plus fusion surgeries in treatment of lumbar degenerative patients [12,13,14,15]. Recent studies reported either similar [16] or slightly different [17] clinical outcomes between the two surgeries. Series of complications have been reported after lumbar surgeries using either techniques that resulted in reoperation of the patients [18, 19]. Reoperation is generally an undesirable outcome, implying persistent symptoms, progression of the underlying diseases, or complications related to the initial operation. The results of reoperation for lumbar degenerative diseases are generally worse than the results of the primary surgery [20, 21]. Patients with one reoperation after lumbar procedures are at considerable risks of further lumbar surgeries [22]. Therefore, reducing or preventing reoperation rate is a primary objective of contemporary spinal surgeries.

Most existing studies of lumbar patients focused on comparisons of clinical outcomes using different surgical techniques, with less focuses on the issue of reoperation [23,24,25,26]. Few studies have reported the timing and causes of reoperation of decompression alone and decompression plus fusion surgeries [27,28,29]. A systematic knowledge is lacking on the aetiology behind the reoperation after the primary surgery. This information is necessary for further improvement in the surgeries for treatment of lumbar patients. Therefore, the aim of this study was to conduct a systematic review of the literature that compared the reoperation rates, timing and causes between decompression alone and decompression plus fusion surgeries. Our null hypothesis was that there is no difference in reoperation rates between these two procedures when used to treat lumbar disease patients.

Methods

A literature search of the following databases (PubMed/MEDLINE, EMBASE and Cochrane Collaboration Library) was performed. Key search terms included “decompression,” “laminectomy,” “discectomy,” “laminotomy,” “laminoplasty,” “fusion,” “fixation,” “instrumentation,” “implantation,” “revision,” “reoperation,” “lumbosacral,” and “lumbar” in different combinations. Inclusion criteria were studies reporting the reoperation rates with underlying causes or risk factors for lumbar decompression alone or plus fusion surgeries. Excluded were non-English articles; in vitro, animal, or cadaveric studies; systematic reviews and meta-analyses; case reports; letters and comments; and studies on paediatric population; studies on trauma, infection, tumour, inflammatory diseases, and deformity; studies involving dynamic stabilization devices or lumbar disc replacement devices; studies with less than twenty patients.

An initial search yielded 6401 articles (Fig. 1). All duplicate publications were excluded. Additional 4147 articles were removed based on the exclusion criteria by analysing their titles. The remaining 550 articles were further filtered by reading their abstracts. In this process, 515 articles were further excluded: 4 for having less than 20 patients; 148 for researching on cervical or thoracic or sacral spine; 363 for not analysing causes or risk factors for reoperation. The remaining 35 publications met all criteria and were included in this systematic review.

Fig. 1
figure 1

Flow chart of literature search process for this systematic review

Full size image

Decompression alone surgery

There are 28 studies discussing the reoperation rates and underlying causes following lumbar decompression alone surgery that are included in this systematic review (Table 1).

Table 1 Characteristics of reoperation after decompression alone surgery
Full size table

Reoperation rate

The reoperation rate is highly variable among different reports. It depends on the length of follow-up time and type of decompression surgeries. In general, the reoperation rate following decompression alone surgery ranges from 1.6 to 41.3% in follow-up times from 3 months to 17.7 years [17, 29,30,31,32,33,34,35,36] (Table 1). Table 2 groups the reported reoperation rates based on follow-up time. Among these studies, the reoperation rates range from 1.6 to 10.8% within 1 year, 6.5–41.3% within 3 years, 3.6–34.0% within 5 years, 4.4–33.8% within 10 years after the primary surgery.

Table 2 Reoperation rates in different follow-up times after decompression alone surgery
Full size table

Kim et al. [37] showed that there were different reoperation rates for different decompression surgeries. In an over 5-year follow-up, the reoperation rates were 18.6, 13.8 and 12.4% for laminectomy, open discectomy and endoscopic discectomy, respectively. However, even for the same type of decompression, the reoperation rate could be different among different studies. Lad et al. [28] reported that the reoperation rate in the patients who underwent a laminectomy alone was 5.7% at 1 year, 10.2% at 2 years and 14.5% at 5 years. However, Ghogawala et al. [38] reported that the reoperation rate reached 29.4% in less than 3 years after the surgery.

The underlying aetiology of reoperation

Same-segment diseases (SSD, including disc herniation, recurrent stenosis, and others) are the major cause of reoperation after decompression alone surgery, accounting for 52.1–100% of the reoperation patients [30, 36, 39, 40]. Leven et al. [41] found that 62% of the reoperations were attributed to the recurrent disc herniation at the same level as in the primary surgery after a standard open discectomy. Similar observations were reported in many other studies [30, 31, 34,35,36, 39, 40, 42]. Recently, Ghogawala et al. [17] reported that reoperations were all at the index levels because of SSD. Besides of the SSD, other causes of reoperation included irrigation and debridement of surgical-site infection, haematoma evacuation, epidural fibrosis, cerebrospinal fluid leakage and so on [31, 34, 35]. In general, SSD accounts for majority of the causes in revision surgeries after decompression alone procedures.

Fusion surgery

There are 19 papers discussing the reoperation rates and underlying causes following fusion surgery that were included in this systematic review (Table 3).

Table 3 Characteristics of reoperation after fusion surgery
Full size table

Reoperation rate

The reoperation rate of fusion surgery is highly affected by the length of follow-up time. From 1 to 11 years of follow-up, the reoperation rates ranged from 0 to 29.3% (Table 3) [16, 17, 28, 29, 43,44,45,46,47,48,49,50,51]. If the patients are grouped by follow-up times, the reoperation rate ranges from 0.0 to 7.0% within 1 year, 7.4–19.6% within 3 years, 7.8–19.3% within 5 years, 3.0–29.3% within 10 years after primary surgery (Table 4). The highest reoperation rate after fusion surgery was reported as 37.5% in a 15-year follow-up [52].

Table 4 Reoperation rates in different follow-up times after fusion surgery
Full size table

The underlying aetiology of reoperation

Symptomatic adjacent-segment degeneration or referred as adjacent-segment diseases (ASDis) [48] in the literature has been reported as a major cause of reoperation after fusion surgery in recent years, with the proportion ranging from 46.9 to 100% of total underlying causes [17, 49, 50]. For example, Irmola et al. [49] and Macki et al. [50] found that the most common pathology leading to reoperation following instrumented lumbar spine fusion was ASDis. Ghogawala et al. [17] recently reported that the reoperations were all at the adjacent levels after fusion surgery. Similar observation was reported at the 15-year follow-up study [52]. Other causes contributing to reoperation included irrigation and debridement for the treatment of surgical-site infection or haematoma, redecompression, screw or rod breakage for implant removal, implant revision for correcting screw position [49, 53, 54].

Comparison of decompression alone and decompression plus fusion surgeries

Among the selected studies in this review, there are 12 papers that directly compared the reoperation rates of decompression alone and decompression plus fusion surgeries (Table 5). Ten of these studies are retrospective and two are prospective. (The prospective study by Irmola et al. [49] only recruited patients with fusion surgery.)

Table 5 Comparison of reoperation rate and aetiology between two surgeries
Full size table

Majority of the studies reported that patients with degenerative lumbar diseases treated with decompression alone tended to have reoperation early [27, 46, 55,56,57], but reoperation rates are similar at longer follow-up time between the two surgeries [16, 28, 44]. Lad et al. [28] revealed a significantly higher reoperation rate in patients who underwent a laminectomy alone than those who had a decompression plus fusion at 1-year follow-up (5.7 vs. 3.3%; p = 0.023); however, at over 5-year follow-up, there were no significant differences between the two groups (17.0 vs. 13.2%; p = 0.347). The prospective study recently published in the New England Journal of Medicine also stated that the percentage of patients who underwent additional operation within a mean follow-up period of 6.5 years was similar between the two surgeries [16].

In a study of patients with grade 1 degenerative spondylolisthesis at an average of 5.9-year follow-up, 10 out of 13 reoperations were due to ASDis in fusion group and 18 out of 25 reoperations were due to SSD in the decompression alone group [29]. In another study of patients with degenerative diseases of lumbar spine, half of the reoperations following a fusion surgery (17/34) were due to ASDis, and 41 out of 72 reoperations after a decompression surgery were at the index levels [58]. In a randomized, controlled trial for grade 1 degenerative spondylolisthesis patients at 4-year follow-up, Ghogawala et al. [17] found that all the reoperations performed in the decompression alone group were at the index levels due to subsequent clinical instability, and in contrast, all the reoperations performed in the fusion group were at the adjacent lumbar level (either disc herniation or clinical instability). In another prospective study for lumbar spinal stenosis at an average follow-up of 6.5 years, Forsth et al. [16] found that 31 patients received reoperations after lumbar fusion with 22 due to ASDis and 3 due to pseudoarthrosis; 28 patients received reoperations after decompression alone surgery with 18 due to restenosis at the index levels and 5 due to ASDis.

In general, these studies reported that the operation rates and clinical outcomes were similar among the patients operated using the two surgeries. ASDis is found to be the major cause of reoperation following the fusion surgeries, while SSD is the major cause of reoperation for decompression alone surgeries.

Reoperation rate with time

A review of the literature indicated that early reported studies showed lower reoperation rate than more recently reported studies [16, 17, 27,28,29, 31, 32, 34,35,36,37, 39,40,41, 43,44,45,46,47,48,49,50,51, 53, 54, 59,60,61,62,63,64,65]. Figure 2 groups the studies reported before 2014 and after 2014 to show the trend of reoperation with follow-up time, i.e., earlier than 2014 and between 2014 and 2017. For treatment of lumbar stenosis patients with the similar age and follow-up period, the reoperation reported by Shabat et al. [61] is 8.7%, which is lower than 23% reported by Forsth et al. [16]. The same phenomenon can also be observed in fusion studies. For treatment of lumbar spondylolisthesis patients with the similar age and follow-up period, Lad et al. [28] reported a 13.2% reoperation rate in 2013, while Gerling et al. [44] reported a 21.2% reoperation rate in 2017. Analysis of these studies implies that the reoperation rate has not been reduced in patients reported in recent years compared to patients reported in earlier years.

Fig. 2
figure 2

Comparison of reoperation rates of lumbar patients after surgery reported before and after the year of 2014: a decompression alone surgery and b decompression plus fusion surgery

Full size image

Discussion

This paper presented a systematic review of published studies on reoperation rates after lumbar surgeries. Specifically, we compare two popular surgical treatments, decompression alone and decompression plus fusion surgeries. Our review indicated that, on average, both surgeries resulted in similar reoperation rates after the primary surgery. However, majority of the reoperations following fusion surgeries are due to ASDis, and following decompression alone surgeries are due to SSD. Reoperation rates were not found to decrease in patients reported more recently than those reported in early times. These findings were largely consistent with our hypothesis that there is no difference between these two procedures in reoperation rate when used to treat lumbar disease patients.

Reoperation after primary surgery could heavily burden healthcare resources. Data from 2002 to 2004 showed that, within 2 years of reoperation, total direct cost of revision surgery was $10,272 for a repeated discectomy [66]. Another study analysed an institutional database from 1997 to 2007 and found that total direct cost of revision discectomy was $39,836 with a 1-year follow-up [67]. The cost for reoperation after fusion is even higher. Sherman et al. [66] reported that within 2 years of reoperation, total direct cost of revision fusion surgery was $27,740. Parker and Adogwa et al. [68] showed that overall mean two-year cost after a revision lumbar fusion was $32,915. Other studies showed that the mean total 2-year costs of revision fusion for pseudoarthrosis, same-level recurrent lumbar stenosis and adjacent-segment disease were $41,631, $49,431 and $47,846, respectively [69,70,71]. Therefore, improvement of surgical techniques to impede reoperation after lumbar surgery is critically important for reducing the healthcare expenditure.

Numerous studies have pursued to identify risk factors for reoperation following decompression alone or plus fusion surgeries with the aim to improve patient outcomes. Majority of the discussion on risk factors for both surgeries are related to patient age, gender, symptoms, coexisting degeneration and type of operation [27,28,29, 31, 32, 35,36,37, 39,40,41, 44, 47, 50, 53, 61, 62, 65, 72, 73]. However, due to different experimental set-up and patient population, inconsistent results have been reported on these risk factors. For example, age has been noted as a risk factor for reoperation but some reported younger ages and others reported older ages as being associated with reoperation [28, 37, 40, 41, 44, 65, 73]. Patients with instrumented arthrodesis have been found to have higher reoperation rates than those without instrumentation at 5-year or longer follow-ups [28]. Facet joint violation (FJV) has been suggested to alter load-bearing capability and to be an independent predictor of undergoing reoperation postoperatively [47]. Besides of the unchangeable inherent factors, smoking cessation [62], losing weight [29, 63], improving comorbidities (especially diabetes) [37, 51] and choosing appropriate operation types and surgical techniques [33, 40] have been reported to be helpful for reducing chances of reoperation after decompression alone surgeries.

For fusion surgery, clarifying the mechanisms of ASDis in order to modify the fusion surgeries has been consistently pursued [74,75,76,77,78]. The aetiology of ASDis has been thought to be multifactorial, stemming from existing spondylosis at adjacent levels [79,80,81], predisposed risk to degenerative changes [82,83,84,85] and altered biomechanical forces near a fusion site [78, 86, 87]. However, scientific data on the exact mechanisms of ASDis after fusion are lacking. Previous studies have mainly focused on the changes in the range of motion (ROM) of adjacent segments, but with inconsistent data reported due to different experiment designs [88, 89]. In addition, the ROMs measured under designed experimental conditions do not represent the intrinsic loading conditions of the spine people experience during activities of daily life [77, 90, 91]. The assumption that preservation of ROM at index level might diminish the incidence of ASDis resulted in the development and application of motion-preserving devices, such as total disc replacement (TDR) and other dynamic devices [92,93,94]. These devices are designed to prevent high stresses applied to the adjacent segment and to avoid the acceleration of structural and mechanical failure in the discs. However, there are insufficient data to support the reduction in ASDis with the use of motion-preserving devices [95,96,97,98]. Radcliff et al. [96] reported no statistically significant differences in reoperation rates at adjacent levels between TDR and fusion patients. No difference in the incidence of ASDis was also reported comparing dynamic devices with fusion surgeries [97, 98]. These results imply that the biomechanical mechanisms of ASDis are complicated and more investigations are warranted.

Recently, changes of disc height in the adjacent level after fusion were assumed as a possible mechanism of ASDis [99]. For example, Kaito et al. [99, 100] reported that increased disc height in the index level after fusion is associated with the development of adjacent-segment degeneration. Since contemporary fusion surgery is aimed to restore normal disc height [101], it could increase the index disc height and consequently increase the pressure at the adjacent segments, resulting in a decrease in the adjacent-segment disc height after fusion. This could lead to higher stresses inside the adjacent discs that could accelerate the disc degeneration. A prospective, longitudinal patient follow-up study is necessary to investigate the disc height changes of both index and adjacent segments that could help reveal the biomechanical mechanisms of ASDis.

It is interesting to find out that the more recently operated patients tend to have higher reoperation rates than patients operated earlier by comparing the data published before and after 2014, despite the advances in surgical techniques, instrumentation and uses of bone growth stimulators as evidenced by the findings that spinal fusion has been greatly improved recently [16, 17, 102]. There is no clear explanation for this observation. A possible reason could be because the threshold for patient reoperation is lower now than before. This observation might also suggest that surgical technique improvements and newer implantation devices do not result in reduced reoperation rate. As the mechanisms associated with the development of ASDis are still not well understood as discussed in this paper, it could be challenging to develop new surgical technologies that can effectively prevent the development of ASDis that leads to reoperation. It is critically important to determine the exact mechanisms that cause ASDis in order to develop new surgical technologies that aimed to delay or prevent reoperation after spinal surgeries.

It should be noted that there are limitations in this systematic review. There is a large variation in experimental conditions among different studies. These include different spine diseases, patient number, patient age, surgical techniques, length of follow-up time, etc. It is difficult to make a consistent comparison of the published data. Therefore, we selected studies for this systematic review using designed inclusion criteria. Further, we discussed the trend of reoperation rate changes with follow-up time. However, there are more studies reporting patient follow-up between 4 and 6 years and less patient reports in shorter or longer follow-up times. Despite these limitations, this systematic review does provide readers on the knowledge of current literature on decompression alone and decompression plus fusion surgeries in treatment of lumbar patients. A prospective, longitudinal study using same patient cohort is necessary for investigation of mechanisms of reoperation after lumbar surgeries.

In conclusion, this review found that the reoperation rates were similar following decompression alone or plus fusion surgeries for treatment of degenerative lumbar diseases. The major cause of reoperation following fusion surgery is ASDis, and the major cause of reoperation following decompression alone surgery is SSD. There is no evidence showing that the reoperation rate has a trend to decline with newer surgical techniques used. The exact mechanisms of reoperation after both surgeries are still unclear. Further researches are warranted to clarify the biological and biomechanical factors that could lead to postoperative reoperations after spine surgeries.