In Sweden, up to 25% of patients report unimproved or worse pain and up to 40% are not happy with the outcome of lumbar fusion [1]. There are many possible reasons for poor results, including instrumentation failure, inadequate surgical technique, and poor patient selection [23]. Other reported reasons for poor surgical results include presurgical pain/function, negative personality traits, emotional status, anxiety/depression, negative outcome expectations, negative coping, and duration of back pain, muscle injury, and workers' compensation [23]. However, the relationship between back muscle injury after surgery and pain remains unclear.

Lumbar surgery induces histological change in muscle in humans. Kawaguchi et al. [4] reported the histological and histochemical changes in back muscle in patients who underwent surgery for lumbar spine disorders for the first time and in patients who underwent repeated posterior lumbar surgery. Histological damage to back muscle because of previous surgical intervention was long lasting. Of 178 patients who had undergone laminectomies 2–5 years earlier, 14 patients had good recovery and 21 patients had poor recovery [5]. Back muscle biopsy was performed in patients with poor outcome. Local denervation, atrophy of paraspinal muscles, and loss of muscular support leads to disability and increased biomechanical strain, and possibly failed back syndrome and pain [5].

The cross-sectional area (CSA) of the back musculature has been measured before and after various types of posterior lumbar surgery in T2-weighted axial magnetic resonance imaging (MRI) using a computer-linked digitizer [6]. The extent of atrophy of the back musculature was determined on MRI after five types of posterior lumbar surgery. However, the clinical outcomes showed no correlation with the extent of atrophy after each surgical procedure [7]. This finding suggested limited application of back muscle atrophy as an indicator of pain.

Back muscle injury and degeneration often occurs after posterior lumbar surgery, and back muscle degeneration may be a cause of back pain after surgery. However, the relationship between back muscle degeneration and back pain remains controversial. In the current study, we aimed to classify back muscle degeneration more precisely using MRI and investigate its relationship with back pain after surgery.

Table 1 shows demographic characteristics of the patients before surgery. Fig. 1 shows a representative MRI for each group. MRI revealed muscle degeneration in all patients after surgery (type 1, 6%; type 2, 82%; and type 3, 12%). Type 2 was significantly more prevalent compared with types 1 and 3 (p<0.01).The average CSA of the multifidus muscle before surgery was not significantly different between the three groups (p>0.05) (Table 2). The average CSA of the multifidus muscle 12 months after surgery was significantly decreased compared with before surgery in all patients (p<0.01) (Table 2). However, there was no significant difference in the average CSA of the multifidus muscle after surgery between the three groups (p>0.05) (Table 2).

Before surgery there was no significant difference in the low back pain score as evaluated by VAS or ODI, or leg pain score evaluated by VAS between the three groups (p>0.05) (Table 3). Low back pain evaluated by VAS and ODI, and leg pain evaluated by VAS significantly improved after surgery (p<0.01) (Table 3). Low back pain and leg pain were not significantly associated with any of the MRI types of muscle degeneration after surgery (p>0.05) (Table 3).

The current study revealed various pathologies of back muscle degeneration after posterior lumbar surgery. Type 2 (fatty) change was most prevalent, and other patients were classified as having type 3 (scar) or type 1 (inflammation or water-like) changes. Low back pain was not associated with any of the MRI types of muscle degeneration after surgery.

Posterior lumbar surgery is one of the main approaches to deal with lumbar spine disorders. Clinical investigations have noted postoperative injury to and atrophy of the multifidus muscle [89] and the extent of atrophy has been significantly associated with postoperative low back pain and functional disability [9]. In the current study, all patients showed muscle degeneration 12 months after surgery. However, low back pain decreased after surgery. In cases of disc herniation or spinal stenosis, we believe that low back pain originates from several structures, including intervertebral discs and spinal nerve roots. Indeed, Toyone et al. [10] reported that in 40 consecutive patients with disc herniation treated by discectomy, all 40 patients were satisfied with the outcome, and suggested that nerve root compression by lumbar disc herniation might be a cause of low back pain. However, patients in the present study still showed slight low back pain 12 months after surgery, and this pain may have originated from degenerated back muscle. The relationship between the extent of back muscle atrophy and postoperative low back pain could not be concluded from the current study.

Presently, type 2 (fatty) muscle degeneration was most prevalent. After posterior surgery in sheep, the main pathological changes noted in the multifidus were the atrophy of muscle bundles, increase in interstices, and decrease in size and number of nuclei [11]. Meanwhile, changes noted on MRI were a decrease in the CSA of the multifidus and infiltration by fibrotic and fatty tissues [11]. Patients with chronic low back pain have a higher percentage of muscle fatty infiltration compared with healthy controls [12]. However, no association between the fat content of the multifidus muscle and pain duration or intensity has been evident [13]. Furthermore, the extent of fatty change in the multifidus muscle is not significantly related with higher ratings of pain intensity [14]. In the current study, low back pain associated with fatty change (type 2) was not significantly different from pain associated with type 1 or 3 atrophy. This finding indicates a lack of association between the fat change in multifidus muscle and postsurgical pain.

Some patients were classified as having type 1 (inflammation or water-like) or type 3 (scar) changes. To our knowledge, there has been no previous report of MRI evaluation of these muscle degeneration changes. Histological findings including muscle degeneration, ballooning of muscle fibers, edema, atrophy, fatty change, and fibrosis and scarring in back muscle from patients who suffered from back pain have been reported [5]. In animal models of posterior lumbar spine surgery, the multifidus muscle has been evaluated histologically and by MRI after surgery [15]. Mainly fatty degeneration occurred. However, necrosis and inflammation and fibrotic change often occur with fatty degeneration [15]. These changes correspond to type 1 and 3 changes seen in the current study.

We wondered whether there is an association between type 1 changes and back pain. We had previously conducted MRI studies of vertebral bone marrow changes in patients with degenerative lumbar disease [16]. Abnormalities associated with decreased signal intensity on T1- weighted spin-echo images and increased signal intensity on T2-weighted spin-echo images (Modic type 1) corresponding to the current type 1 change correlated with segmental hypermobility and low back pain [16]. We have also reported that Modic type 1 endplate abnormalities in discogenic pain patients are related to inflammation and axonal growth into the abnormal bone marrow induced by cytokines such as tumor necrosis factor-alpha [17]. However, in the current study, low back pain associated with type 1 change was not significantly different from that associated with type 2 and 3 changes. Further study is needed to clarify the relationship between back pain and the type of muscle degeneration.

The current study has some limitations. First, it is a small-sized retrospective study and the number of patients was limited. Second, we used 1.5 T MRI, which limited resolution. Third, the patient population only included those with lumbar spinal stenosis. As mentioned above, there is the possibility that back pain originated from spinal nerves, other structures, and other muscles, such as the longissimus and the iliocostalis. Fourth, we did not examinethe relationship of fatty infiltration to back muscle and back pain before and after surgery. Fatty infiltration to back muscle has been reported to be crucial factor to back pain [18]. Therefore we could not evaluate pain that was purely associated with postsurgical muscle degeneration.

Various pathologies of back muscle degeneration were evident after posterior lumbar surgery. Fatty change was most prevalent, and other patients had scarring or inflammatory-like changes. However, no type of back muscle degeneration was correlated with back pain after surgery.