Surgical treatment of trigeminal neuralgia. Results from the use of glycerol injection, microvascular decompression, and rhizotomia

The study aims to assess the efficacy and safety of surgical treatment of trigeminal neuralgia (TN) in our department and to identify prognostic factors. Seventy patients receiving surgical treatment for TN during the period 2003-2004 were included in this retrospective study. The surgical procedures used were glycerol injection (GI), microvascular decompression (MVD), or rhizotomia (RIZ). All patients were divided into spontaneous onset TN type1 (brief lancinating pain) or TN type 2 (continuous pain component). Two patients had bilateral TN; each side was regarded as a separate case. These 70 patients had a total of 160 interventions (110 GI, 40 MVD, and ten RIZ) performed in the period 1998-2007. Data were obtained by chart review and telephone interview. Patients provided information about pre- and postoperative pain characteristics including subtype, duration, intensity, and the use of antiepileptic drugs. Outcome was evaluated using a pain vector diagram. To quantify self-reported pain, we developed a new vector-based pain diagram. The subtype of TN was shown to be a very important prognostic factor. One year after MVD, 90% of patients with type 1 TN still had positive effect, whereas this was only true in 73% of patients with type 2 TN. After RIZ, the results were 71% and 33% for types 1 and 2, respectively. For comparison, GI had a significant lower effect but if the treatment led to hypoesthesia, 41% continued to have a positive effect 1 year after surgery, compared to only 24% if postoperative sensation was normal. Type 2 TN was found to be dominated by women with left-sided TN outside the V2 dermatome and with a lower probability of a neurovascular conflict. As expected, 1/5 of the cases developed postoperative hypoesthesia in the face following a nerve destructive procedure (RIZ and GI). Using MVD, the risk of serious side effects was about 4%. Complementary and alternative treatment had no general or permanent effect in the investigated population-quite the contrary. Regarding prognosis and outcome, we find that it is very important to classify TN in subgroups (types 1 and 2). Dealing with medically treatment-resistant type 1 TN, MVD and RIZ are reasonably safe and effective interventions. The surgical results dealing with type 2 TN are still very poor. All patients with medically treatment-resistant TN should be offered referral to a neurosurgical unit with experience in treating this painful disease. We recommend using a vector-based pain diagram when evaluating the outcome of multiple interventions.