Axial pain with upper or lower limb pain due to spinal diseases is very prevalent debilitating problem and epidural steroid injection (ESI) is one of most popularly used conservative treatment for spinal pain. But concerns for adverse effects associated with steroid usually used in ESI are widely spread among not only patients but even also physicians, which is the main reason of reluctance to perform ESI. This frequently acts as a barrier, which leads the patients, which could have been treated by injection, to be irreversible chronic pain conditions or undergo more invasive treatment.

In the aspect of ESI, utilized steroids are usually categorized into particulate and non-particulate steroids. Particulate steroids, with ester group, has the property of producing particles, aggregates, or packings when they or their preparation mixed with local anesthetics were injected. They can cause brain or spinal cord infarct by vascular occlusions, as well, they yield prolonged clinical efficacy by being deposited into the target tissues longer. As well, prolonged maintenance in tissues produces accumulation and further systemic side effects, even though steroid is locally injected. Thus, exact knowledges about the characteristics according to different steroid types is necessary to make decision appropriate use of steroid in epidural injection. Aside from innate property of steroid, consideration of dose, number of injections, and intervals between injections should be taken in order to promote clinical effects and lessening the adverse effects. Understandings of these aspects pave the way to eliminate exaggerated or sometimes mythical fear or concerns, which further helps to establish appropriate treatment method to the patients with spinal pain.

This article reviewed characteristics of steroid and proper therapeutic strategy in ESI such as dose, repetition, or intervals in the aspect of benefit and hazard, on the basis of literatures that had been published.

The common synthetic corticosteroids used in epidural injection are derivatives of prednisolone. All have anti-inflammatory potencies per dose unit somewhat greater than that of cortisol. Methylprednisolone is the methyl derivative of prednisolone, whereas betamethasone, dexamethasone, and triamcinolone are all fluorinated derivatives of prednisolone [1]. Additionally, betamethasone is an isomer of dexamethasone.

Particulate steroids, usually existing as acetate salt form with ester group includes methylprednisolone acetate (MPA), triamcinolone acetonide (TA), and betamethasone acetate (BA). These required a process of hydrolysis by cellular esterase to release active metabolites in order to produce clinical effects, eventually showing delayed onset of therapeutic effects. They are less soluble in a hydrophilic environment and form microcrystalline suspensions due to the presence of corticosteroid ester. They are composed of relatively larger particulate size than non-ester forms. When injected into tissues, they have property of forming aggregates and being densely packed, which lead to longer duration of action by tissue deposit of large and slowly absorbing particles [1].

While, non-particulate steroids with non-ester group include dexamethasone sodium phosphate (DSP) and betamethasone sodium phosphate (BSP). They, which come in sodium salt form, have more water-soluble characteristics than ester group. They, formed as clear solution rather than suspension, can be taken up rapidly by cells, featuring early onset of action. As well, they or their mixture with local anesthetics have small particle size, forms less aggregate, and are less packed than ester group, allowing shorter duration of action. An important exception is the potential for dexamethasone to unexpectedly form a crystalline suspension when mixed with the local anesthetic ropivacaine [2].

Dexamethasone palmitate (DEP) is a derivative of dexamethasone, consisting of dexamethasone and a fat emulsion containing lecithin. The C16 chain in DEP enhances its hydrophobicity, facilitating strong interactions with phospholipid aliphatic chains. This property enables DEP to be taken efficiently by macrophages and ensures a robust distribution in inflammatory tissues. In addition, DEP lowers the probability of forming particles and eventually reduces the risk of vascular adverse events more profoundly than any other non-particulates, when mixed with local anesthetics [3,4].

Celestone Soluspan (BSP-BA) has two components at the same time, combination of betamethasone ester and betamethasone salt. This has quick onset and long-acting property. Celestone Soluspan contains both soluble and insoluble forms of betamethasone in a single preparation for injection and can be considered as particulate in overall nature [5]. The characteristics of steroids used in ESI are presented in Table 1.

Table 1 . General characteristics of steroid used in epidural injection.

	Betamethasone sodiumphosphate and betamethasone acetate	Methylprednisolone acetate	Triamcinolone acetonide	Dexamethasone sodiumphosphate	Betamethasone sodium phosphate
Dosage form	Suspension	Suspension	Suspension	Solution	Solution
Benzyl alcohol	No	Yes	Yes	Yes	No
Equivalent dose	0.75	4	4	0.75	0.75
Particle, aggregation, packing	Variable	Large	Large	Small	Small
Solubility	Acetate form, insoluble; sodium phosphate form, freely soluble	0.001	0.0002	Soluble	Soluble

Ester group, commonly called as particulate steroids are expectedly more associated with vascular side effects due to its characteristics of tendency of large particle and aggregation as mentioned above section, when they are intravascularly introduced. This produces the concern for devastating side effects such as brainstem or upper spinal cord infarct due to embolus or thrombosis resulted from intraarterial needle penetration and occlusion or aggregated drug particles.

The main blood supply to the spinal cord is derived from the spinal branch of a local artery that then divides into anterior and posterior radicular arteries. The spinal cord and nerve roots are supplied by radicular arteries which travel at each of the vertebral levels through the neuroforamen. Cervical spinal cord is supplied by corresponding radicular arteries originated from the vertebral, deep cervical, and ascending cervical artery. At the thoracolumbar regions, radicular arteries are branched from the intercostal and lumbar arteries. The artery of Adamkiewicz is one of lumbar anterior radicular arteries, usually observed at the left side of T8 level or variably between T9 and L2, which contributes to blood supply to large area covering the lower thoracic and upper lumbar spinal cord [5-7].

The particles or aggregations formed by steroid accidentally intravascular administered play a role in producing embolic occlusion of radicular artery. Tiso, et al. demonstrated the microscopic appearances of various steroid particles. Particles in DSP and BSP appeared to be lucent rodlike, while particles of MPA and TA tended to be opaque and amorphous. Over time, particles of MPA and TA gradually coalesced into large aggregates and produced significant amounts of larger particles. This might represent a factor contributing to microvascular “sludging” and subsequent occlusion/infarction [5].

Derby, et al. presented similar results. While, microscopic findings of preparation of DSP mixed with lidocaine showed only small particle and no aggregates formation, TA showed large particle and aggregate formation. Another particulate steroid, MPA preparation formed small sized particles and these particles were densely packed. BSP-BA produced large aggregates by various sized particles. This study suggested that DSP was less likely to cause arterial or capillary obstruction if inadvertently injected into a vertebral or foraminal artery and physicians should consider utilization of non-particulate steroid especially when performing cervical transforaminal ESI [8].

Orduña-Valls, et al. measured the size and numbers of aggregates of three steroids preparations. Aggregates present in TA and BSP-BA were statistically larger than in DSP and moreover, TA produced significantly larger aggregates than BSP-BA five minutes after mixing. The results demonstrate that TA and BSP-BA preparation, especially TA has the possibility to occlude the vessels when intravascular administered [9].

Transforaminal approach is more associated with embolic infarct from particulate formation than interlaminar or caudal approach because radicular artery is nearly located at needle trajectory. In transforaminal approach, cervical is most prevalent, thoracic is lesser, and lumbosacral is least because cervical foramen is most narrow so as to be vulnerable to vascular invasion during needle advancement. Although cervical and thoracic transforaminal injections consist of less than 5%, they occupy about 99% of serious neurological complication [10].

Arachnoiditis is another neural side effects, although its incidence is very low. The possible mechanism is that inflammatory reaction on pia-arachnoid membrane, when steroid is intrathecally injected, cause intrathecal scarring, clumping, and tethering of neural tissue. Intrathecal injection of large volume of steroid, especially particulate steroid, is assumed to be predisposing to arachnoiditis. But the relationship of particulate steroid with arachnoiditis is questionable so that technical consideration, accidental intrathecal injection is more determinant factor than steroid property [7].

Demyelination and neural degeneration are other neural side effects that can be caused by steroid. This is results of neurotoxicity by the preservative or drug vehicle contained in the steroid formulation. Benzyl alcohol included in MPA, TA, and DSP preparation, not in BSP-BA is regarded to be caudal factor [1].

Steroid administered for epidural injections can produce systemic side effects due to systemic absorption of steroid from epidural space. Among them, hyperglycemia and blood pressure elevation are immediate side effects. Most notable long term side effect is hypothalamic pituitary axis (HPA) suppression. In patients receiving lumbar ESI, one study found that 20.3% of subjects had greater than 50% decrease from baseline of cortisol level at three weeks. But this side effect was only observed in those treated with MPA or TA but not DSP or BSP [11]. Other studies about intraarticular steroid injection, the time interval from HPA suppression to normal range was TA was longer than BSP [12,13]. These results suggested that particulate steroid was more closely associated with HPA suppression than non-particulate steroid, as expected by the fact that particulate steroid had the property of prolonged tissue deposits.

The studies comparing the duration of HPA suppression between two different doses found that the mean time to have return to baseline HPA function was 19.9 ± 6.8 days after a single injection of 40 mg TA and 8.0 ± 2.4 days after 20 mg TA [14,15]. Eighty-six percent of patients demonstrated cortisol suppression one week after ESI with 80 mg MPA while 53% of those who received 40 mg MPA showed cortisol suppression, but no significant difference was found at 3 weeks between two groups [16]. Ultimately, the evidence supporting dose response effect as to HPA suppression is not confirmative yet.

According to Lee, et al.’s study, significant HPA suppression was not correlated with glucocorticoid equivalent dose and there was no significant difference as to number of steroid injections, steroid types, and injection locations between significant and non-significant HPA suppression group, which was observed both cases such as serum cortisol withdrawn within 6 weeks as well as between 6 weeks and 12 months after the most recent injection. Only more recent and repetitive injection with particulate steroid showed trends toward HPA suppression [17].

It is not evident that HPA suppression is significantly associated with steroid dose or injection numbers. No set number of injections nor a specific threshold was not presented [18]. But high number of injections during same period is possible factor to HPA axis suppression. Empirically, 2-3 weeks intervals between injections are recommended to prevent cumulative HPA axis suppression effects, especially when particulate steroids are used. Low dose of non-particulate steroids at appropriate spacing is favored to avoid accumulation of steroid and further prolonged HPA suppression, if higher dose or particulate steroid cannot significantly guarantee to facilitate clinical improvement.

Bone mineral loss or osteoporosis is one of long-term systemic side effects related to steroid. Significant reductions in bone mineral density (BMD) were associated with a cumulative MPA dose of 200 mg over a one-year period and 400 mg over three years, but not associated with less than 200 mg of MPA for postmenopausal women or less than 3 g for healthy men [19]. This was comparable to another study’s results that identified approximately 14 ESIs with a cumulative TA dose of approximately 400 mg for treatment of low back pain could reduce BMD in postmenopausal women [20]. But some studies failed in demonstrating such a relationship, with one report stating that the risk of osteoporotic compression fracture decreased with each additional ESI. Consequently, it was concluded that limited evidence of complications of osteoporotic compression fracture might be more likely to occur after 14 or more ESI exposures [18].

Wong, et al.’s study showed that lumbar transforaminal ESI (TFESI) with 3 mg of BSP obtained not inferior in reduction of Numeric rating scale (NRS) and opioid consumption to lumbar TFESI with 6 mg of BSP [21]. Similarly, there was no significant difference of Visual analogue scale (VAS) and Oswestry disability index among the groups of TFESI using 4 mg, 8 mg, and 12 mg of DSP [22].

Also, in the studies of lumbar ESI using particulate steroid, comparable results were found. Lumbar ESI with 40 mg of MPA did not reveal significantly different VAS reduction from 80 mg of MPA at 2 weeks and 3 months [23]. A 40 mg of TA, while this caused higher blood sugar level during 3 days after lumbar ESI, did not achieve significantly more pain reduction than 20 mg of TA [24].

These results postulated that higher dose of steroid was not useful despite more concerns for side effects. If higher dose may not significantly facilitate clinical improvement, low dose of steroid is favored.

Repetitive injections at 2 weeks intervals were recommended to fulfill enough clinical response when only partial response was obtained by first injection. Among 51 patients undergoing TFESI, 20 (40%) and 9 (18%) patients needed 2 and 3 sessions to achieve satisfactory results. Even four injections were required for enough response in one (2%) patient [25]. Because the patients receiving clustered injections achieved statistically significant cumulative benefit, repeat TFESI was necessary to provoke clinical outcomes in those with incomplete response at 2 weeks postinjection [26].

Lee, et al.’s study supported the importance of planned repeat ESI in the patients with partial pain reduction after first injection. This study compared the clinical results between the group of repeat ESI at a prescribed interval of 2 to 3 weeks (group A) and the other group of intermittent ESI performed only based on patients’ need (group B) among partial responders after first injection. Group A could obtain significantly longer duration of satisfactory pain control and the longer time until reinjections than group B with the similar number of injections needed [27]. Overall, repeat injections at planned intervals about 2-3 weeks allowed for enough clinical outcomes by accumulating effects before abolishing previously obtained effects by first injection.

There has been many literatures investigating whether particulate steroid with an ability of high tissue deposit and its consequence, long duration effect, can accomplish significantly better clinical efficacy than non-particulate steroid. The study comparing clinical efficacy between DSP 7.5 mg, or with TA 40 mg at 1 month showed significantly more pain reduction but no significant difference in functional score [28]. But large amount of comparison studies found that non-particulates accomplished non-inferior or even superior clinical outcomes as to pain control and functional improvements [29-34]. A retrospective study of lumbar TFESI for radicular pain identified that DSP was non-inferior to the particulate steroids in categorial data such as ≥ 50% reduction in NRS and ≥ 40% reduction in Roland-Morris score (RMS) at 2 months and was even superior in terms of continuous data such as mean NRS and RMS at 2 months [29]. A double blind randomized controlled trial (RCT) of lumbar TFESI demonstrated that DSP obtained reasonably similar effectiveness to TA until 6 months, although DSP group received slightly more injections than the TA group to achieve the same outcomes [30]. According to another RCT of lumbar TFESI, DSP showed no significant difference in pain reduction and functional outcomes at 3 months and even better functional improvement at 6 month than BSP-BA by multivariate regression analysis [31]. The studies about cervical TFESI for radicular pain with relatively short term follow up period, 1 month, also revealed no significant difference between particulate and non-particulate steroids in clinical efficacy [32-34]. Meta-analysis also concluded that because particulates was not found to be significantly better in pain control than non-particulates, in consideration of concerns for the safety profile of particulates, it might be prudent to switch to non particulates [35].

ESI has been used for controlling the axial or limb pain secondary to various spinal pathologies including disc herniation and stenosis and validated as useful conservative methods through many clinical studies and reviews supported by moderate or strong or moderate evidence level with relatively low risks [36,37]. Furthermore, ESI can provide the patients with opportunity to avoid extensive surgical treatments that can be more associated with serious complications [38]. But to facilitate the clinical efficacy of ESI in addition to minimize potential adverse events, several recommendations were suggested.

For prevention of vascular adverse effects such as intravascular occlusion and furthermore, brain and spinal cord infarct, non-particulate steroids are recommended and furthermore their non-inferior clinical efficacy to particulate steroids is strongly supportive to preference for non-particulates. In detail, it is recommended that particulates should not be used in cervical TFESI. In lumbar TFESI, particulates are not recommended as routine use and at the level above L3, but can be allowed when their uses are needed. For prevention of systemic side effects, especially of HPA suppression, non-particulates are favored. High dose, short interval, and large number of injections might increase risk of systemic side effects, but conclusive evidence or guideline as to appropriate number, intervals, and dose does not exist [18,39]. For a single level lumbar ESI, NASS recommends no more than 3 injections within 6 months and maximum 6 injections per year. Spacing ESIs at least three weeks apart to allow for recovery after potential HPA suppression may limit long-term sequelae of cumulative steroid exposure [18]. When only partial and unsatisfactory response is obtained by first injection, repeat injections are required to fulfill more complete and prolonged clinical effects by accumulating treatment effects without concerns of overtreatment or abuse [27]. High dose of steroid is not necessary because no evidence is found that high dose has the ability to promote better outcomes in comparison with low dose [22-24].

While ESI is popularly used treatment which is relatively less invasive than surgery, the exaggerated fear or concerns for steroid frequently render patients not to take injection and therefore to lose the chance of recovery. The exact and appropriate understanding for steroid used in ESI is inevitable to maximize the clinical benefits as well as to avoid adverse effects, which contributes to provide the patients with undergoing the optimal treatment and to avoid chronic pain conditions or unnecessary more invasive treatment sometimes associated with harmful results. Based on many literatures, pre-planned ESIs at appropriate intervals using non-particulates and low dose of steroid is enough to obtain the treatment goal and to avoid unwanted side effects.

No potential conflict of interest relevant to this article was reported.