Sensitivities of noninvasive tests for central nervous system vasculitis: a comparison of lumbar puncture, computed tomography, and magnetic resonance imaging.

OBJECTIVE

To determine the sensitivities of lumbar puncture (LP), computed tomography (CT), and magnetic resonance imaging (MRI) for central nervous system (CNS) vasculitis.

METHODS

We reviewed the charts and radiologic studies of 20 patients with angiogram positive CNS vasculitis.

RESULTS

The patients ranged in age from 7-72 years, with a mean of 43.7 +/- 15.5 years. Twelve patients (60%) were women; 8 (40%) were men. Seven had primary angiitis of the CNS (PACNS), 8 had rheumatological illnesses (4 had systemic lupus erythematosus), and 5 had vasculitis due to other diverse etiologies, including 2 who had CNS infections. LP was positive in 8/15 angiographically proven cases, for a sensitivity of 53% [95% confidence interval (CI): 27-79]. The sensitivity of CT was 65% (11/17) (95% CI: 38-86), and that of MRI 75% (12/16) (CI: 48-93). The use of LP plus either CT or MRI was more sensitive than LP alone: for LP and CT, the sensitivity was 92% (11/12) (CI: 62-100), and for LP and MRI, 100% (12/12) (CI: 74-100). The combination of CT and MRI was not more sensitive than either test alone.

CONCLUSION

The sensitivities of LP, CT, and MRI for angiogram positive vasculitis are only modest. CT and MRI may be completely normal in cases of CNS vasculitis when the angiogram is positive. LP and either CT or MRI are of additive benefit in the diagnosis of CNS vasculitis, and should be done before angiography; and, in a patient with both a normal LP and a normal CT or MRI, a CNS angiogram is unlikely to be positive for vasculitis.