Salt jab not 'a miracle cure' for back pain

"Relief for millions as revolutionary new saline jab spells end of agonising back pain,” is the sensationalist claim in the Daily Express. The actual truth of the matter is that the study in question only provided very limited evidence.

In some people with lower back pain a herniated (‘slipped’) disc causes compression or irritation of the nerve roots as they leave the spinal cord and this causes pain going down into the legs (neuropathic pain). Sciatica is the common term for this type of pain.

Sometimes epidural injections (where drugs are injected into the outer most part of the spinal canal) of steroid or local anaesthetic may be used to try and relieve neuropathic pain.

In trials assessing epidural steroid injections for low back pain, an epidural injection of saline/salt solution is often used as a placebo control.

The main aim of this study was to see if this “placebo” injection might itself have an effect on back pain compared to another type of control – an injection not given into the epidural space, such as into the muscles.

Evidence gathered by the researchers, which was of variable study design and quality, suggested that epidural saline injection may be more likely to produce a positive response than non-epidural injection. However, the difference in pain score was small and not significant.

It is difficult to know what to usefully conclude from this study. It demonstrates that one type of placebo (a saline epidural injection) is more effective than another type of placebo (an injection that doesn’t go into the epidural space), but it doesn’t show that epidural saline injections are as or more effective than the conventional low back pain treatments.

Where did the story come from?

The study was carried out by researchers from Johns Hopkins School of Medicine and other US institutions and was funded by the Centers for Rehabilitation Sciences Research, Uniformed Services University of the Health Sciences, Bethesda, Maryland.

The study was published in the peer-reviewed journal Pain Medicine.

The standard of the Express’ reporting of this study was poor. Excitable claims that this research would lead to a “simple injection [that] could banish back pain for good among millions of sufferers” cannot be supported by the evidence provided by the study. The researchers themselves concede that the evidence only suggests that “epidural nonsteroid injections may confer some benefit”.

The Mail Online’s coverage was slightly more restrained, but again, made it appear that the study shows that an epidural injection of salt solution to back pain may be more effective than current treatments. This is not the case.

What kind of research was this?

This was a systematic review which aimed to investigate whether epidural injections of salt solution might have an effect on back pain. These injections are often used as a “placebo” treatment in randomised controlled trials (RCTs) of steroid injections, but the researchers wondered if they might actually be having a beneficial effect on back pain themselves.

In the UK, the treatment of low back pain normally involves exercise and movement, and short-term treatment with anti-inflammatory tablets such as ibuprofen. Various other non-invasive therapies may also be tried. Some people who have persistent low back pain also get pain going into their legs that comes from the point where the nerves come out of the spinal cord.

Sometimes epidural injections of anaesthetic or steroid may be used to try and relieve the pain in these people. Epidural injections are injections into the epidural space – the area within the spinal vertebrae but outside of the spinal cord which consists of the nerves. Epidural injections are used to numb the nerves in this area to relieve pain.

As mentioned above, in an RCT, the placebo version of a steroid epidural injection that is often used is an inactive salt/saline solution. An alternative placebo is a non-epidural injection, for example, injection given into a muscle (intramuscularly) rather than given in the epidural space. This injection could be either saline or steroid. The current study aimed to review the literature to compare how effective the two “placebos” – epidural non-steroid injections and non-epidural injections – were at providing back pain relief.

Because both of these injections are considered to be placebos, few trials have bothered to compare them directly. This meant that the researchers had to use a technique called indirect meta-analysis to compare them.

Rather than pooling the results of studies that compared (x) treatment with (y) treatment as in a conventional meta-analysis, an indirect meta-analysis is more complicated – it looks at trials comparing (x) and (y) with a third treatment (z).

The results of these trials are then used to estimate the difference that would be expected if (x) was directly compared with (y).

The reason this convoluted approach was used was due to the lack of evidence of direct comparisons, in RCTs, between epidural non-steroid injections and non-epidural injections

Direct RCT comparisons are a better way to compare treatments, but when these aren’t available, or there are very few of them, this method allows researchers to make comparisons they wouldn’t otherwise be able to make. For the results to be a good estimate of the difference between two treatments, the different RCTs being analysed need to be in the same types of people.

What did the research involve?

The researchers searched two literature databases to identify RCTs conducted in adults with low back pain (with or without additional nerve pain extending into their legs) and where:

  • a treatment group received epidural injections with steroids (or another drug aiming to relieve pain)
  • a control/placebo group received an epidural injection of a non-active solution (such as saline), or a non-epidural injection (where an injection was given into the muscle rather than into the epidural space)
  • data on the participants’ outcomes was collected up to 12 weeks after the final injection

Studies were assessed for quality and those that provided numerical pain data (such as pain ratings) were pooled in the meta-analysis. The main outcomes they were interested in were:

  • treatment response/success (compared with non-success)
  • pain reduction on a rating scale

What were the basic results?

The researchers identified 43 eligible trials comparing epidural steroid injections with a control injection. Over half the studies (65%) were considered to be of high quality. The individual trials included between 22 and 228 people. The exact position, number, frequency and dose of the injections used in the trials varied.

They found the following results:

  • 35 studies compared epidural steroid injections (active treatment) with epidural non-steroid (control) injections and around a quarter of these studies (23%) found that the active treatment was more effective than control
  • 12 studies compared epidural injections with the control of non-epidural (intramuscular) injections and over half of these (58%) found that the epidural injections were more effective than non-epidural injections
  • only three small studies (309 participants in total) directly compared non-steroid (saline) epidural injections and non-epidural (intramuscular) injections with other active treatments. However as the researchers note, none of these studies had set out to find the difference between the two different control treatments. None of these studies found significant differences between the controls
  • the indirect meta-analysis of response to treatment involved 23 studies (1,512 people) comparing epidural steroid injections with epidural non-steroid (saline) injections, and seven studies (663 people) comparing epidural injections with non-epidural injections. It found that a person was twice as likely to have a positive response with epidural non-steroid injections, than non-epidural injections (relative risk 2.17, 95% confidence interval [CI] 1.87 to 2.53)
  • the indirect meta-analysis of pain score, involved 22 studies (1,936 people) comparing epidural steroid injections with epidural non-steroid (saline) injections, and four studies (619 people) comparing epidural injections with non-epidural injections. It found a small but non-significant difference between epidural non-steroid injections and non-epidural injections, again marginally in favour of the epidural non-steroid injections (mean score difference -0.15 points, 95% CI -0.55 to +0.25)

How did the researchers interpret the results?

The researchers conclude that: “Epidural non-steroid injections may provide improved benefit compared with non-epidural injections on some measures, though few, low-quality studies directly compared controlled treatments, and only short-term outcomes (less than 12 weeks) were examined.


This review primarily aimed to look at whether the epidural saline injections that are used as an inactive “control” in trials of epidural steroid injections might actually themselves have a beneficial effect on back pain. To do this it compared their effect against another commonly used inactive “control” – non-epidural (intramuscular) injections of either steroid or saline.

An effective non-steroidal type of injection would be welcome as steroid injections can only be given on an infrequent basis due to the risk of side effects. The treatment is also not suitable for some groups of patients, such as those with liver disease.

Sadly the results of this study do not show that saline injections are an effective alternative to current treatments.

The main findings were that:

  • Only three small studies were available that had directly compared epidural saline injections with non-epidural injections. These studies found no difference between these two treatments.
  • Trials of epidural steroid injections using epidural non-steroid (saline) injections as a control were less likely to show an effect of the epidural steroid injections on back pain than trials using non-epidural (intramuscular) injections as a control.
  • In indirect meta-analysis epidural saline injections appeared to be more effective than non-epidural injections (either non-epidural saline or steroid).

This provides some evidence that what have been considered to be “placebo” epidural injections of salt solution may have more effect on back pain than placebo injections given somewhere other than the epidural space. When looking at the size of the difference between the groups on pain ratings, this difference was very small, and not big enough to be confident that it didn’t occur by chance. 

The fact that the analysis used indirect comparisons means that conclusions need to be made more cautiously than if they came from trials directly comparing these injections.

Any differences between the trials in the types of people they included and their methods could make the results less reliable.

Ideally these results would be confirmed by trials directly comparing the injections if researchers feel that they are strong enough to warrant further assessment. It seems unlikely that doctors will change how they treat back pain based on these results. Epidural injections of any kind carry their own risks, and are only used on selected patients.

Also, it’s important to remember that this study does not tell us how the epidural saline injections compare to other conventional treatments for back pain, such as epidural steroid or anaesthetic injections.

There are also a range of non-drug treatments for back pain, including exercise and physiotherapy.

NHS Attribution