David Seminowicz, University of Maryland
Baliki and colleagues’ newest study is huge: huge in that they scanned 56 participants four times each over one year (plus an extra 13 patients in a validation group); and huge in terms of the results and the potential clinical implications. The study included 17 healthy controls and 39 subacute back pain (SBP) patients. After dividing the subacute group into those who developed chronic back pain (SBPp) and those who recovered (SBPr), the authors found that only the SBPp group had decreased brain gray matter volume in bilateral insula and nucleus accumbens (NAc), and left S1 at later time points. Even more striking was that the resting state connectivity between NAc and medial prefrontal cortex (MPFC) at the first visit scan predicted whether a patient’s back pain would persist at the 1-year follow-up.
What do the findings mean and what are the clinical and research implications of this study? It is very often unclear why, given two people with similar acute injuries, one might go on to develop chronic pain, and the other might recover completely. The answer might not be in changes in the periphery or spinal cord, but in the brain, and this study suggests that the MPFC-NAc circuit might be crucial. For research, it means we need to keep exploring brain mechanisms of pain at all disease time points. Neuroimaging studies have typically focused on acute pain in healthy controls or in people with chronic pain, but little attention has been given to the subacute stage. Clinically, the findings in this study suggest that we might be closer than we’ve ever been to being able to identify a person likely to develop chronic pain, and treatment plans for those individuals should start early on, hopefully preventing the onset of chronic pain.
A potentially important finding (see supplementary info) was the other predictors of SBPp/SBPr group membership: pain duration was a significant predictor on its own and second in predictive value to MPFC-NAc connectivity, while sensory pain (on the MPQ) and early drug use both approached significance. Early drug use was defined as starting drugs before the first scanning session. This suggests that early (and likely aggressive) treatment after injury may help protect against the development of chronic pain. Including MPFC-NAc connectivity, MPQ sensory, and early drug use in a multiple regression model gave the best predictive value of all.
So, an immediate clinical message that might be taken from this study: if you hurt your back, do not hesitate to treat the pain. You might be doing more than relieving the pain, but actually changing your brain to prevent the onset of a more serious and hard-to-treat disease.
A finding of such huge mechanistic and clinical consequences begs for replication, but this study is likely never to be replicated: it’s too big, too expensive. One part of it, however, could be replicated with some ease: scanning patients at the (sub)acute stage, then calling them a year later to find out if they have developed chronic pain. In fact, Baliki et al. did such a validation test by scanning 13 additional patients. Again, MPFC-NAc connectivity was the best predictor of pain at 1-year follow-up.
The data presented are part of an ongoing project, so we can expect to see more fascinating results coming out of this study. The Apkarian lab shares its data on an ftp portal, so hopefully other researchers will have an opportunity to mine this massive dataset for other gems. It would be interesting to see if a model-free, whole brain approach (such as that used in Wager TD, et al., J Neurosci. 2011 Jan 12;31(2):439-52) could discriminate SBPr and SBPp groups, and whether MPFC-NAc connectivity would be an essential part of the predictive pattern, or if some new predictive brain pattern would emerge.
How does PRF find papers of the week? See our PubMed search string and complete search results by clicking here.
What are you reading? Please share your favorite topical search string and we'll post it here.
Copyright © 2013 Pain Research Forum
All site content, except where otherwise noted, is licensed under a Creative Commons BY-NC-ND License. Additional permissions for appropriate reuse available on request.