Glycerol rhizotomy versus gamma knife radiosurgery for the treatment of trigeminal neuralgia: an analysis of patients treated at one institution.

BACKGROUND: Medically refractory trigeminal neuralgia (TN) has been treated with a variety of minimally invasive techniques, all of which have been compared with microvascular decompression. For patients not considered good surgical candidates, percutaneous retrogasserian glycerol rhizotomy (GR) and gamma knife (GK) radiosurgery are two minimally invasive techniques in common practice worldwide and used routinely at Jefferson Hospital for Neuroscience. Using a common pain scale outcomes questionnaire, we sought to analyze efficacies and morbidities of both treatments. METHODS AND MATERIALS: Between June 1994 and December 2002, 79 patients were treated with GR and 109 patients underwent GK for the treatment of TN. GR was performed with fluoroscopic guidance as an overnight inpatient procedure. GK was performed using a single 4-mm shot positioned at the root exit zone of the trigeminal nerve. Radiation doses of 70-90 Gy prescribed to the 100% isodose line were used. Treatment outcomes including pain response, pain recurrence, treatment failure, treatment-related side effects, and overall patient satisfaction with GK and GR were compared using a common outcomes scale. Using the Barrow Neurologic Institute pain scale, patients were asked to define their level of pain both before and after treatment: I, no pain and no pain medication required; I, occasional pain not requiring medication; IIIa, no pain and pain medication used; IIIb, some pain adequately controlled with medication; IV, some pain not adequately controlled with medication; and V, severe pain with no relief with medication. We used posttreatment scores of I, II, IIIa, and IIIb to identify treatment success, whereas scores of IV and V were considered treatment failure. Results were compiled from respondents and analyzed using SAS software. Statistical comparisons used log-rank test, univariate and multivariate logistic regression, Fisher's exact test, and Wilcoxon test with significance established at p < 0.05.
RESULTS: There were 63 evaluable GK patients and 36 evaluable GR patients. The median follow-up time was 34 and 29 months for the GR and GK groups, respectively. The median age was 69 and 70 years and the median number of years with TN pain was 6 and 7 years in the GR and GK groups, respectively. Thirty-one GR (86%) and 58 GK (92%) patients achieved a successful treatment outcome (p = 0.49). The median time to pain relief was < or = 24 h in the GR group and 3 weeks in the GK group (p < 0.001, ordinal logistic regression). Nineteen GR and 26 GK patients experienced pain recurrence or pain never relieved (p = 0.30). The median time to pain recurrence was 5 and 8 months in the GR and GK groups, respectively (p = 0.22). At last follow-up, however, a statistically significant greater number of GR vs. GK patients had failed treatment. Twelve or 33% of GR patients, whereas 8 or 13% of GK patients, had BNI scores of 4 or 5 (p = 0.019, Fisher's exact test). When both initial and late treatment failures were combined, the overall rate of treatment failures was 39% in the GR group and 24% in the GK group (p = 0.023, log-rank). More GR patients developed facial numbness and facial numbness morbidity than GK patients: 19 GR (54%) and 17 GK patients (30%) developed new facial numbness and 12 GR and 7 GK patients reported either somewhat or very bothersome facial numbness (p = 0.018). On multivariate analysis, only treatment with GK and severity of pain before treatment significantly predicted treatment response. GK patients were more likely to have a lower pain score at last follow-up than were GR patients (p = 0.005, OR = 4.3), and patients with pretreatment pain scores of 5 were more likely to have lower posttreatment pain scores than patients with pretreatment pain scores of 4 and lower (p = 0.003, OR = 4.0).
CONCLUSION: Despite greater facial numbness morbidity and a higher failure rate, GR provided more immediate pain relief than GK. GR therefore should be considered in patients with disabling trigeminal pain requiring urgent pain relief. For all other patients, GK provides better long-term pain relief with less treatment-related morbidity, and should therefore be considered the preferred treatment for patients with medically refractory trigeminal neuralgia who are not suitable candidates for microvascular nerve decompression.