Lumbar radiculopathy can be well-managed conservatively in the primary health care setting, but many patients with persistent disabling radicular pain need attention in a specialty clinic. The majority of patients first attend a public specialty clinic in Hong Kong with pain chronicity of more than 12 months, as they have had no significant clinical response to conservative management in primary health care, private medical specialists, traditional Chinese medicine, or alternative medicine, and they anticipate a long waiting time in a public hospital. Epidural steroid injection is commonly practised by orthopaedic surgeons, neurosurgeons, rehabilitation specialists, pain specialists, and interventional radiologists worldwide. The thresholds for offering epidural steroid injection by clinicians and acceptance by patients are variable, however.

Transforaminal epidural steroid injection (TFESI) is one of the more common approaches of epidural steroid injection, along with the interlaminar and caudal approaches. The technique is target-specific and the best route for delivering medication to the ventral epidural space (Fig 1a) and dorsal root ganglion,1 where most pathological changes occur.2 3 The least amount of drug with a relatively higher drug concentration is required to reach the primary site of pathology compared with interlaminar and caudal epidural steroid injections.4

Transforaminal epidural steroid injection is a useful procedure for lumbar radiculopathy.5 The technique provides neural blockade to anaesthetise the target nerve root for diagnostic purposes, and interrupts nociceptive input and self-sustaining activity of the neurons. Steroid provides anti-inflammatory effect (inhibition of pro-inflammatory synthesis and release of mediators) and produces longer-term pain relief, primarily for radiculopathy. The prognostic value of TFESI for surgical outcomes has been reported, with better surgical outcome in TFESI responders with chronic lumbar radiculopathy than in non-responders.6 The technique, however, does not alter the ultimate need for surgery.7

Underlying sepsis, malignant disease, and coagulopathy are considered to be contra-indications for spinal injection. The perceived benefits and threshold of offering TFESI as an adjunct to conservative treatment for patients with lumbar radiculopathy attributed to disc herniation or spinal stenosis are variable among orthopaedic surgeons, rehabilitation specialists, and pain specialists. The objective of this study was to identify the diagnostic, therapeutic, and prognostic values of TFESI for lumbar radiculopathy in Chinese patients in Hong Kong.

Patients were placed in the prone position on a radiolucent operating table. A 22-G spinal needle was inserted into the target neuroforamen with fluoroscopic image guidance. The target nerve root and its epidural space were outlined by water-soluble non-ionic contrast, ensuring epidural flow of contrast with no intravascular, intradural, or subcutaneous infiltration (Fig 1). A mixture of 1-mL methylprednisolone acetate 40 mg and 1-mL bupivacaine 0.5% was injected. Finally, the intact spinal needle was removed.

All patients who received TFESI for lumbar radiculopathy at Alice Ho Miu Ling Nethersole Hospital, Hong Kong from 1 January 2007 to 31 December 2011 (5 years) were identified by the electronic medical record system in the hospital. Retrospective review of all the medical records identified patients with numeric pain rating scale score (NPRS) of 4 to 7 (out of 10) who took less than three types of analgesics for at least 8 weeks as conservative treatment, or patients with persistent disabling pain with NPRS of >7 despite taking more than three types of analgesics for at least 1 week as conservative treatment. Patients with NPRS of <4, missing record of post-procedure response, or who had received previous lumbar spinal injection and lumbar spinal surgery were excluded.

The first part of the study assessed the response rate to diagnostic block by the local anaesthetic effect of TFESI for all patients with a clinical diagnosis of lumbar radiculopathy attributed to disc herniation or spinal stenosis, with compatible magnetic resonance imaging (MRI) findings of laterality and affected level. The second part of the study assessed the therapeutic and prognostic values of the steroid effect of TFESI as an adjunct to conservative treatment prior to assessment of surgical need.

The threshold for surgery for patients with lumbar radiculopathy attributed to prolapsed intervertebral disc (PID) and spinal stenosis, in general, considered factors of disabling pain resulting in inability to meet activity demands, clinical MRI findings compatible for laterality and site of compression, and medical fitness for general anaesthesia and major spinal surgery. Spinal fusion might be considered for patients with concomitant spondylolisthesis with instability or anticipated instability resulting from optimal surgical decompression in the lateral recess or foraminal stenosis and concomitant disabling discogenic low back pain that has not responded to conservative treatment.

An immediate response on the procedure day was considered to have a positive diagnostic value. Patients who reported pain reduction of greater than 50% at the first follow-up visit 4 weeks after TFESI were considered to have a positive therapeutic response to the steroid effects. The response duration, proportion of patients finally requiring surgery, whether decompression alone or spinal fusion besides decompression was needed, and risk factors that affected the response to TFESI were retrospectively reviewed.

Comparisons were carried out for all patients, as well as patients with PID or spinal stenosis only. Associations between responses to TFESI and risk factors for symptom chronicity and industrial injury were done by Fisher’s exact test or Mann-Whitney U test where appropriate. Non-parametric tests were done because some continuous variables were not normally distributed. The Statistical Package for the Social Sciences Windows version 20.0 (SPSS Inc, Chicago [IL], US) was used for all statistical analysis. A two-sided P value of ≤0.05 was considered statistically significant.

A total of 241 patients were recruited into this study. Nine patients were excluded for the following reasons: TFESI responders with PID and L5 radiculopathy had had symptom duration of less than 1 week and TFESI was not considered to be adequate first-line conservative treatment (n=3); and TFESI immediate responders did not return for first follow-up (n=2) and TFESI responses were not documented in the medical records (n=4) so the response durations for these six patients could not be verified. Therefore, the total number of eligible participants was 232 (110 men and 122 women; mean age ± standard deviation [SD]: 55.6 ± 14.3 years). The mean age of patients with PID and spinal stenosis were 37.4 ± 7.5 years and 60.3 ± 11.7 years, respectively. The symptom chronicity ranged from 8 days to 23 years with a median of 12.0 months, as well as 25th, 75th, and 90th percentiles being 7.0, 36.0, and 60.0 months, respectively in patients with lumbar radiculopathy. No statistically significant difference (P=0.402) in symptom chronicity between the PID and spinal stenosis groups was noted (median [interquartile range] duration: PID group 12.0 [8.0-24.0] months vs spinal stenosis group 12.0 [6.0-36.0] months). Fewer PID patients (n=48; 20.7%) needed TFESI than spinal stenosis patients (n=184; 79.3%) with lumbar radiculopathy in the study period.

L5 was the most commonly affected level of radiculopathy (n=150; 64.7%) regardless of whether a patient had single or multiple levels or underlying pathology of PID or spinal stenosis (Table 1). Therefore, post-ganglionic block of the L5 nerve root by L5-S1 TFESI was most commonly done.

The immediate response rate to TFESI was 80.2% in 186 patients with clinically diagnosed lumbar radiculopathy and MRI of the lumbar spine suggesting nerve root compression. Overall, 218 (94.0%) patients were affected at a single level and 14 (6.0%) were affected at multiple levels (Table 1). The immediate response rates to TFESI were 175 (80.3%) in the single-level radiculopathy group and 11 (78.6%) in the multiple-level radiculopathy group. There was no statistically significant difference in the immediate responder rate between patients with PID or spinal stenosis (P=0.877). No complications were reported.

The final need for surgery of TFESI immediate responders was noted in 10/39 (25.6%) patients in the PID group and 43/147 (29.3%) patients in the spinal stenosis group (Table 2).

Of the 232 patients, 106 (45.7%) were offered surgery, of whom 65 (61.3%) accepted surgery. The mean time from TFESI to uptake of surgery was 7.9 months. There was a statistically significant shorter median time to definitive surgery in the PID group (10.0 months) than in the spinal stenosis group (19.2 months) [P<0.01]. This reflects the fact that PID patients with failed first-line conservative treatment who needed TFESI for lumbar radiculopathy were likely to accept surgery earlier than patients with spinal stenosis. Patients with PID were younger (mean age, 37.0 years) than spinal stenosis patients (mean age, 59.2 years) undergoing surgery, which might be related to less daily activity demand, higher perceived operative risks, and older people being more psychologically reluctant to undergo surgery.

Of the 65 surgical patients, 23 (35.4%) underwent decompression surgery alone, with a mean time from TFESI of 5.45 months (SD, 5.25 months; median, 3.6 months; range, 8 days to 17.63 months). The remaining 42 (64.6%) patients required spinal fusion in addition to decompression surgery, with a mean time from TFESI of 9.37 months (SD, 7.23 months; median, 7.22 months; range, 14 days to 25.33 months). The difference in time to surgery for these two groups was statistically significant (P=0.012). More patients with a short-term response to TFESI underwent surgery (Fig 2) and TFESI was commonly used as a preoperative assessment tool.

The analgesic effect of TFESI lasted for less than 1 week (poor response) in 68 (29.3%) patients, for 1 to <3 weeks (short term) in 35 (15.1%) patients, for 3 to 12 weeks (intermediate) in 37 (15.9%) patients, and for more than 12 weeks (long term) in 92 (39.7%) patients. More patients with spinal stenosis underwent surgery in the short-term pain reduction group (1-<3 weeks), and the association between response to TFESI and surgery for spinal stenosis was significant (P<0.01), but no significance was noted for PID patients (P=0.067) [Table 3].

Poor response (mean, 34.3 ± 50.9 months) to TFESI (no immediate response and pain reduction duration of <1 week) was significantly associated with chronicity of symptoms (vs 23.1 ± 28.3 months in patients with positive response) [P=0.047]. Pain reduction duration had no significant association with symptom chronicity for pain reduction of less than 3 months and 3 months or more in the PID (P=0.225) and spinal stenosis (P=0.250) groups (Table 4).