BACKGROUND: Patients without genuine zygapophysial joint pain exhibit a positive response to uncontrolled diagnostic blocks. We determined the false positive rate of single uncontrolled diagnostic zygapophysial joint block in patients with chronic back pain. METHOD: Thirty-five patients with chronic lower back pain were studied prospectively. All patients underwent a series of medial branch of dorsal rami blocks with lidocaine and bupivacaine starting at the symptomatic level to a maximum of three levels or until the pain was abolished. RESULTS: All patients proceeded with the injections. 16 patients (45%) had a definite or greater response to the initial lidocaine injection at one or more levels but only 5 (14%) had a 50% or greater response to a confirmatory injection of 0.5% bupivacaine. Using the confirmatory blocks as the criterion standard, the false positive rate of uncontrolled diagnostic blocks was 37% and the positive predictive value was only 31%. CONCLUSION: Uncontrolled diagnostic blocks may be unreliable for the diagnosis of lumbar zygapophysial joint pain.