JC: Turn it Down to 11 – Benzos for Back Pain

The topic of lower back pain of undifferentiated origin is one that is close to my heart for at least two reasons. The first one is professional, as daily we see and manage a significant number of patients with lower back pain presenting to the ED where I practice. It is never easy to manage these patients as the pain can be anything from mild and uncomfortable to severe and frankly debilitating.

The second reason is purely personal as I suffered a lumbar disc prolapse a couple of years ago, possibly secondary to poor technique at the gym. I was in significant pain for several months and this had direct repercussions on my life and particularly my enjoyment of sporting activities.

Regardless of where in the world you practise EM, I have no doubt that you’ll have come across patients presenting with low back pains, and the question of what analgesia works best (and let’s face what we mean by best is an outcome we struggle to define) has come up during your practice. There certainly seems to be a wide variation in drugs prescribed in the ED where I work ranging from simple paracetamol, to NSAIDs and mild opiates (Ed – although not in the same way as seems to be the case in the US), sometimes combined with some benzodiazepines.

As for the evidence behind benzodiazepines in acute lower back pain? Well, it has been rather scant. I was therefore pleased to come across this paper published in the Annals of Emergency Medicine1 looking at the efficacy of oral diazepam when added to naproxen. Let us dissect this paper but remember to read the full article before drawing any conclusions…

What was the purpose of the trial?

The purpose is clearly defined at the beginning of the paper: the authors set out to evaluate whether combining a benzodiazepine (namely diazepam) with an NSAID (naproxen) is more efficacious than NSAID monotherapy for the treatment of acute, non-traumatic and non-radicular lower back pain. The population is clearly defined. Maybe too well even, but I will come back to that a bit later…

What type of trial is this?

This was a randomised, double blind, comparative ED-based efficacy trial conducted in two urban EDs.

What did they do?

Patients were enrolled during an ED visit for acute musculoskeletal back pain and followed up by telephone at seven days, then again at three months after discharge. The follow-up periods were arbitrarily defined and one could argue that a follow-up further down the line would have been beneficial as non-traumatic lower back pain is often a chronic condition. The enrolment was continuous 24/7 during the study period of nine months which is remarkable as this is difficult to achieve in many centres.

Tell me about the patients

The patients were adults aged between 21 to 69 years (a completely arbitrarily chosen age range which could be open to criticism) who presented with low back pain defined as non-traumatic and non-radicular. Only patients who were discharged home were included and those who had a score greater than 5 on the Roland-Morris Disability Questionnaire. I had not come across this questionnaire before but it is a validated 24-item tool used to evaluate functional impairment secondary to low back pain. You can read more about this here if you are also unfamiliar with it.

A reasonable list of exclusion criteria can be found in the original paper.

545 patients were approached, 418 excluded and therefore randomising 116 in total in equal numbers to placebo and diazepam groups. The baseline characteristics were comparable between groups and the primary analysis was intention to treat.

Interventions and outcomes

Patients were randomised in blocks of 4 according to a generated random-number sequence (read more about block randomisation here). Diazepam and placebo were placed into identical capsules as part of blinding . Despite the unusual dose and frequency of diazepam/naproxen (twice daily as required), the randomisation process looked robust enough to me for the authors to look at this question.

I have never taken benzodiazepines in my life but wonder if the patient could be truly blinded as I suspect one can potentially feel the pharmacological effects of diazepam (notably somnolence).

The primary outcome for the study was improvement on the above-mentioned Roland Morris Disability Questionnaire between ED discharge and day 7 follow up. A 5 point improvement on this scale is generally considered a clinically significant one and the authors provide a link for us to have a look.

Half a dozen secondary outcomes were also looked at and I would suggest that you have a look at the original paper should you be interested in these. The length of this blog post does not allow me to look at every single of them in detail but they seemed to be all clinically relevant to me.


One week after ED visit, the diazepam group improved by a mean of 11 (95% CI 9 – 13) on the Roland Morris Disability Questionnaire, whereas the placebo patients improved by… 11 (95% CI 8 -13) (no typo here, they observed the same difference in both groups). The 95% CI for mean difference was therefore 0.3: -2.8 to 3.5.

The authors highlight that the between-group difference achieved neither clinical nor statistical significance. I think this is an important point to make as some papers present results as being statistically significant but this does not always translate into any relevant clinical benefit for the patients (the so-called clinical significance our patients are interested in).

I would urge you once again to read the paper in full to have a look at the other half dozen secondary outcomes (these included use of rescue medication, side effects etc.).


The authors quite rightly discuss the issues around looking at only a subset of patients with lower back pain resulting in a  supra-selected population. The majority of patients I see and treat during my practice almost invariably present with radicular symptoms, probably a worrying feature that makes them seek medical help. Issues around external validity derived from the fact that the study was based in a socioeconomically deprived area, the “as needed” schedule of medication regime, the absence of an assessment for muscle spasm at point of recruitment are also discussed by the authors.

I would like to add to the above the presence of a number of confounders for which the study was not adjusted. The most important that springs to mind is any type of debilitating pain is accompanied by a certain level of anxiety. Could diazepam have acted as an anxiolytic rather than muscle spasm relaxant in these patients, knowing that pain has a psychological component attached to it? Also, patients were taught alternative techniques like stretching, exercises, cold/hot packs etc at point of discharge in an uncontrolled fashion and this could have altered outcomes in both groups.


The authors concluded that diazepam did not appear to confer any benefit (at one week or three months after ED discharge) beyond that of placebo when added to naproxen for the treatment of non-traumatic low back pain.

So what?

This is an interesting paper that looked a relevant ED topic (Ed – very relevant to us in Virchester see posts here2and bestbets here3) but with the above-mentioned limitations. I personally would be interested in reading a paper including patients with radicular symptoms, a more robust analgesia regime and frequency as I am not sure this paper completely and reliably answered my question…

But what conclusion(s) do YOU draw for your practice?




Before you go please don’t forget to…

Friedman BW, Irizarry E, Solorzano C, et al. Diazepam Is No Better Than Placebo When Added to Naproxen for Acute Low Back Pain. Annals of Emergency Medicine. February 2017. doi: 10.1016/j.annemergmed.2016.10.002
Back Pain. St.Emlyns. http://stemlynsblog.org/tag/back-pain/. Published 2017. Accessed February 24, 2017.
Back Pain. BestBets. http://bestbets.org/database/browse-by-topic.php?CategoryID=114. Published 2017. Accessed February 24, 2017. [Source]

1 Comment

  1. Rory

    1 week ok. What about pain at 8hrs; 24hrs? I don’t expect people to be taking Benzo or orphenadrine at a week anyway.


Thanks so much for following. Viva la #FOAMed

Translate »