Back surgery poser
There
are bad days, and then there are days when someone asks you to pick the bones
out of a randomised trial of back surgery compared with usual care. What we
want is a large, high quality randomised trial, which gives a clear answer.
Some of us might be backing surgery, others analgesics and sympathy, but all of
us want the answer. Two studies, one a randomised trial [1], the other a cohort
of patients refusing to participate [2] do little but confuse us more.
Trial
In
all 1,244 patients with lumbar intervertebral disk herniation were enrolled out
of 1,991 eligible (747 refused to participate). Of these, 501 were enrolled
into a randomised trial, and 743 in an observational study (which included
surgery and usual care. Follow up was for two years.
The problem
In
the randomised trial, not everyone in the surgery group had surgery, and quite
a lot of those not assigned to surgery had it anyway. The percentage in each
group having surgery was quite similar, as Figure 1 shows. Not surprisingly,
the outcomes in both groups were much the same.
Figure 1: Actual surgery in RCT groups
Most
of the observational cohort had surgery (528/743), and did better than those
who did not. In fact, in all four groups, the only consistent difference was
that those not randomised and not having surgery did less well than all others.
What does it mean?
The
papers come with erudite editorials, providing lots of perspective. Perhaps the
only suggestion is that it will need a sham surgery trial to sort this lot out,
but that may run into ethical storms. The one approach not mooted is the large,
prospective, high quality, registry study, which might help us figure which
patients do best from back surgery.
References:
- JN Weinstein. Surgical vs nonoperative treatment for lumbar disk herniation: the spine patient outcome research trial (SPORT) - a randomized trial. JAMA 2006 296: 2441-2450.
- JN Weinstein. Surgical vs nonoperative treatment for lumbar disk herniation: the spine patient outcome research trial (SPORT) observational cohort. JAMA 2006 296: 2451-2459.
previous story