??prednisone for acute sciatica

JAMA just had an article on the utility of oral steroids in patients with acute sciatica (seelbp disc steroids jama 2015 in dropbox, or JAMA. 2015;313(19):1915-1923). The goal of this study was to see if short-term oral prednisone was effective in improving disability and pain, as well as decrease the need for invasive procedures. Details:

--randomized controlled trial in Kaiser California comprising 269 adults (mean age 46, 55% male, baseline Oswestry Disability Index ODI  score of 51, with 100 being the worst, and pain score of 6.7, with 10 = worst). All had radicular pain < 3 months (mean 30 days), and  a herniated lumbar disc documented on MRI. Prednisone was given as 60 mg x 5 d, then 40 mg x 5 d, than 20 mg x 5 d.

--primary outcome was ODI score change at 3 weeks; secondary outcome was ODI change at 1 year, change in lower extremity pain, spine surgery and Short Form 36 Health Survey (SF-36) Physical and Mental Component summaries. [note: there is no clear consensus on what is a clinically significant change in ODI score. Estimates are in the 5-15 range. They arbitrarily chose 7 as an important change]

--results:

                --ODI scores changed dramatically in each group by 3 weeks, but more so in those on prednisone (ODI went from 51.1 to 37.5 in the placebo group and 51.2 to 32.2 with prednisone, which is less than their own 7 point mark of clinical significance, though did achieve statistical significance at p=0.006). [this dramatic improvement even with placebo reflects the fact that acute sciatica does usually resolve on its own over time]

                --over the 52-week period, there was no further benefit of prednisone in ODI at the 6 week, 12 week, or 24 week checks, but did reach significance (for unclear reasons) at 52 weeks, with difference of 7.4 points

                --no significant difference in pain scores throughout

                --the SF-36, which was scored 0-100, improved in those on prednisone a tad in the Physical Component at 3 weeks (3.3 point difference, p=0.001), though not at 52 weeks. No difference in  Mental Component at 3 weeks but 3.6 point difference at 52 weeks (p=0.02).

                --no difference in surgery rates at 52 weeks

                --adverse effects more common in prednisone group (49.2% vs 23.9%, p<0>001), with the usual culprits: insomnia, nervousness, increased appetite. [I should point out that there was a pretty strong placebo component as seen in these adverse effects, and the reality is that those on prednisone know more often that they are on an active drug – in this study 75% on prednisone thought they were on prednisone, whereas 53% on placebo thought they had an active drug]

So, seems like a very small and questionably significant clinical benefit with prednisone, as reflected both in the small absolute change in the ODI and the fact that the ODI difference was only significant at 2 of the 5 points it was measured, though it happens to be that those 2 points were the ones they used for their formal outcome measurements……  And, this benefit really could have been a placebo benefit, given that 50% more patients thought they were on active drug than placebo. I just want to reiterate and generalize this point: not all RCTs are the same. With many drugs, the patients can tell if they are on an active medication by the adverse effects, and in those cases we should be very careful in blindly attributing benefit to the drug (ie, if the patients really think they are taking an active drug, they may be more suggestable that the drug is helping).  Bottom line: hard for me to justify using a 15 day course of prednisone, a powerful drug with such minimal clinical benefits, if any. And I have certainly seen patients who develop acute psychotic reactions to prednisone, though in their study, the adverse effects were not so bad.