Risks of the tests are minimal.  Myelography requires a spinal tap, so a risk of spinal headache is most frequent but usually clears with bed rest and increased liquid intake.  CT and MRI scans have negligible risk. However, there is also risk of interpreting the results improperly. If considered in isolation, abnormalities of the lumbar spine detected by MRI, CT or myelogram examination can be meaningless.

There is a high prevalence of abnormalities in the lumbar spine on MRI examination of people without back pain. In one study of patients without back pain, only 36 percent had a normal disk at all levels. About half had a bulge of at least one intervertebral disk, and about 20 percent  had at least one disk protrusion. Therefore, with the high prevalence of back pain in the population, the discovery of a bulge or protrusion on an MRI scan in a patient with low back pain is frequently coincidental.

A study using CT to examine 52 people without symptoms found the prevalence of herniated disks to be 19.5 percent in people under the age of 40 years and 26.9 percent in those over the age of 40.  We know that the incidence of disk protrusions increases with age, but most studies use all ages to determine the false-positive rate of imaging.

In another study, 24 percent of 300 myelograms in people without symptoms showed abnormalities of the lumbar disk.  So there is no question that all three imaging modalities can give false-positive results. Any of these imaging studies may also fail detect a herniated disc when one is present, giving a false-negative result.  The MRI and CT are better at detecting HNP than is myelography, but no imaging study is perfect. The false-negative rates are hard to know because the true presence of a HNP can only be determined at surgery, and even surgeons sometimes aren’t sure of what they found.

I think the data used in LOBAK from my study17, in which a non-operating surgeon observed the operation and recorded whether a HNP was actually found, is still valid.  At that time, only myelography was available and many patients with significant neurological findings were operated on (“explored”) despite a normal myelogram.  Out of 106 patients operated on after a myelogram and who did have a HNP, the myelogram had been normal in 26 – a false-negative rate of 25%. Fortunately, the CT and MRI scans are much better at detecting HNP, with as false-negative rate of only 8-10%.