By Angela Mailis-Gagnon MD, MSc, FRCPC (PhysMed)
The three-phase bone scan has been used since the mid-1970s to diagnose CRPS. An intravenous(IV) injection of a particular radiolabelled substance that has a special tendency to concentrate in the bones is administered and a technician takes images of the body part in question, looking for the initial phase of “blood flow.” Immediately, he will look again for the second phase of “blood pool.” Finally, approximately 2 hours later, images will show the concentration of the radiolabelled material in the actual bones; this is the “delayed phase.”
There has been a characteristic pattern of activity in the involved limb, which in the early 1980s was described as “pathognomonic” -a sign or symptom upon which a diagnosis can be made- of CRPS. However, these were retrospective studies, which looked at patients who first had the bone scan for suspected CRPS and then the data were analyzed. These studies could not tell how many of these patients who had a limb trauma, but no CRPS symptoms, had an abnormal bone scan as well.
CRPS has been reported fairly frequently after a fracture; however, in one study mentioned in the recent volume of Progress in Pain Research and Management1, only 16 percent of the patients diagnosed with CRPS 8 weeks after trauma had the characteristic bone scan pattern.
A meta-analysis of 19 published papers related to CRPS combined the data from all these studies and used special calculations to give an overall idea how well abnormalities in the scan correspond to those who have clinical evidence of CRPS. In only half the cases, the bone scan pattern was pathognomonic of CRPS.1 Furthermore, as the disease progresses, the changes in the bone scan go away.
Interestingly, about 12 years ago, a patient of mine who had all the symptoms and signs of CRPS I had a surgical sympathectomy (considered an appropriate type of surgery for the syndrome at that time). Her three-phase bone scan before the surgery was absolutely normal; after the procedure, the radiologist told me that, according to the bone scan pattern, she had “florid CRPS.” Intrigued by this, we followed a number of patients before and after sympathectomy.2 and proved that sympathectomy itself produces a pathognomonic CRPS bone scan.
In my line of work, CRPS I is a common diagnosis, representing 14 percent of all patients with neuropathic pain referred to my clinic (unpublished data from a very recent analysis of 784 consecutive patients who attended my program). Frequently patients are referred to me because of a “CRPS bone scan.” Many of these people don’t have symptoms any more, but the bone scan is still active. Other patients are sent to me because of an abnormal bone scan (done in the process of regular follow-up), but they have no other symptoms. “CRPS bone scan” can occur also in the healthy leg, not the one with the CRPS! Obviously this is due to the abnormal and excessive weight bearing demands on the good limb when the patient favors the leg with symptoms. In other cases, I have seen a few patients with factitious disorder imitating CRPS (these emotionally-disturbed individuals may ligate the arm or the leg creating the picture of CRPS), and some have a “CRPS bone scan,” which becomes normal when they stop ligating the limb. Finally, I have a significant number of patients with florid symptoms and signs of CRPS and a negative or unclear three-phase bone scan.
On the other hand, several studies have shown that typical CRPS bone scans can be seen in patients with diabetes and diabetic neuropathy.3
In my view, the value of a three-phase bone scan in the diagnosis of CRPS is very limited. A three-phase bone scan neither makes nor excludes the diagnosis of CRPS by itself. I always teach my students: “Never treat a test result, treat the patient.” In other words, a pathognomonic “CRPS bone scan” can confirm a diagnosis of CRPS only when my patient has the clinical signs and symptoms of CRPS.
1. Progress in Pain Research and Management , CRPS/RSD: Current diagnosis and management, pg 152-154) published by the International Association for the Study of Pain, 2005,
2. Mailis A et al. Alterations of the three phase bone scan after sympathectomy. Clin J Pain, 10:146-155, 1994.
3. Mailis A. Is Diabetic Autonomic Neuropathy protective again Reflex Sympathetic Dystrophy? Clin J Pain, 1995;11:76-84).
Angela Mailis-Gagnon MD, MSc, FRCPC (PhysMed) is the Medical Director, Comprehensive Pain Program, Toronto Western Hospital, Canada.