What works in back pain?
Clinical bottom line
Muscle relaxants are probably useful in acute low back pain, and analgesics and NSAIDs in chronic and probably acute low back pain.
Reference
LA Machado et al. Analgesic effects of treatments for non-specific low back pain: a meta-analysis of placebo-controlled randomized trials. Rheumatology 2009 48: 520-527.
Study
This systematic review used a search strategy recommended by the Cochrane Back Review Group, and searched five electronic databases for placebo controlled randomised trials of interventions for acute or chronic low back pain. Placebo could include sham interventions, or minimal interventions. Trials with specific causes of back pain, or with surgery, within the previous 12 months, were excluded.
Pain outcomes had to be continuous, but any duration of therapy was allowed.
Results
There were 76 trials for 34 different interventions (Table 1). Few interventions were reported in three trials or more, with more than 500 patients, and for longer than a few weeks. Only for analgesics, herbal medicines, muscle relaxants, and NSAIDs did we have more than 500 patients. For the rest of the interventions, only three had as many as 250 patients, with seven over 200 and 23 below 200 patients.
For the four interventions with adequate data:
- Analgesics produced significantly less pain than placebo in three trials of chronic back pain over 10 weeks.
- Herbal medicines (which not stated) produced significantly less pain than placebo in four trials of chronic back pain, but only over three weeks.
- Muscle relaxants produced significantly less pain than placebo in nine trials, all but one of acute back pain over one week.
- NSAIDs produced significantly less pain than placebo in seven trials of acute and chronic back pain over an average of six weeks.
For no other intervention was there sufficient evidence to tell whether they worked or did not work.
Table 1: Interventions for low back pain
Number of |
|||
Intervention |
Trials |
Patients |
Mean duration (weeks) |
| Acupuncture | 4 |
149 |
3 |
| Analgesics | 3 |
748 |
10 |
| Anticonvulsants | 1 |
96 |
10 |
| Antidepressants | 4 |
217 |
7 |
| ATP | 1 |
161 |
4 |
| Back school | 1 |
26 |
10 |
| Behaviour therapies | 2 |
34 |
6 |
| Colchicine | 1 |
15 |
12 |
| Elecrtroacupuncture | 1 |
25 |
2 |
| Exercise | 3 |
204 |
4 |
| Facet injections | 3 |
257 |
3 |
| Heat wrap therapy | 2 |
255 |
1 |
| Herbal medicines | 4 |
705 |
3 |
| Immunoglobulins | 1 |
41 |
2 |
| Infra red | 1 |
38 |
7 |
| Isis | 1 |
116 |
52 |
| Laser | 2 |
76 |
4 |
| Magnets | 1 |
36 |
3 |
| Massage | 1 |
51 |
4 |
| Muscle relaxants | 9 |
820 |
1 |
| Nerve blocks | 1 |
17 |
2 |
| Neuroreflexology | 1 |
70 |
0 |
| NDMA antagonists | 1 |
43 |
8 |
| NSAIDs | 7 |
1349 |
6 |
| Physiotherapy | 1 |
120 |
4 |
| Proiotherapy | 3 |
263 |
13 |
| PTIT | 3 |
139 |
19 |
| Radiotherapy | 1 |
32 |
6 |
| RF denervation | 4 |
223 |
7 |
| Shortwave | 1 |
65 |
4 |
| SMT | 6 |
247 |
2 |
| TENS | 4 |
178 |
2 |
| Traction | 1 |
150 |
5 |
| Vitamin B12 | 1 |
60 |
2 |
(and for those who ask, we don't know what all the interventions are, either)
Comment
The simple fact is that for most interventions there is just not enough data on which to make an informed opinion. And this systematic review is probably putting more of a gloss on the data we have than is probably justified, because it used a non-standard quality scoring scale which had 11 items, and was almost certainly insensitive, letting through trials likely to be biased in one important way or another. So this is all we have in the way of randomised trials.
What we can probably say is that muscle relaxants are probably useful in acute low back pain, and analgesics and NSAIDs in chronic and probably acute low back pain. For herbal medicines we have to reserve judgement because the trials were too short to be valid for chronic low back pain, because we don't know what the herbal medicines were, and because herbal medicines are too often adulterated, particularly with NSAIDs.
It may be possible to go one step further, if only we could get better insight into some of the RCTs of NSAIDs and coxibs in back pain. There is a suggestion - but only that - that at least half of people can get a reasonable response - equivalent to a 30% or 50% reduction in pain.