Trigeminal neuralgia (TN) is a chronic pain condition that is difficult to control with medical management. ^1,2 Carbamazepine is the recommended first-line drug (level A) for pain management in TN; however, patients often require additional agents such as oxcarbazepine, gabapentin, or baclofen, among others. ³ Disabling side effects are common, with at least one adverse event reported in 66% of patients taking carbamazepine. ⁴ Furthermore, since the efficacy of medical management typically decreases over time, many patients turn to surgical intervention for more durable pain relief. ⁵ Surgical options include microvascular decompression (MVD), stereotactic radiosurgery (SRS), partial sensory neurectomy, and percutaneous procedures (glycerol rhizotomy, radiofrequency ablation, and balloon compression). ⁶

MVD, when successful, is known to provide the longest period of pain relief. The hallmark study by Barker et al. showed that, after MVD, 70% of patients had pain- and medication-free status for 10 years, and the annual rate of pain recurrence was less than 1%. ⁷ Although MVD is effective when feasible, patient selection is critical for its success, and patients are still exposed to the operative risks associated with open cranial surgery. Additionally, many TN patients do not meet surgical criteria for MVD, such as those who lack evidence of vascular compression on imaging. Finally, some patients prefer less invasive interventions, such as SRS.

In this study, the authors conducted a multicenter retrospective review of patients diagnosed with type I TN at 12 international Gamma Knife centers. All patients in this study (n = 404) had no prior surgical intervention or irradiation. Indications for SRS included medication intolerance, failure of medical management, or both. The purpose of this study was to determine how SRS affects reliance on pain medication regimens and to evaluate pain outcomes based on the time interval between diagnosis and SRS.

Methods

Under IRB approval, an international group of 12 Gamma Knife centers retrospectively reviewed patients diagnosed with type I TN who underwent SRS as their primary surgical intervention. Patient demographics, disease characteristics, treatment plans, medication histories, and relevant outcomes (with an emphasis on pain outcomes) were reviewed and documented for analysis.

Results

Time Course to Pain Relief

The median time to pain relief after SRS for the whole cohort was 27 days (IQR 21–30 days) with a 78.6% pain relief rate at 90 days (Fig. 1 left). When stratified by time to SRS from diagnosis, patients who received SRS within 4 years of initial TN diagnosis achieved significantly faster pain relief than those who received SRS beyond 4 years (median 21 days [IQR 15–30 days] vs 30 days [IQR 23–31 days]; HR = 1.23 favoring ≤ 4 years, p = 0.041, unadjusted) (Fig. 1 right). The 90-day pain relief rate for those who received SRS ≤ 4 years from diagnosis was 83.8% versus 73.7% in patients who received SRS > 4 years from diagnosis. After adjusting for the only other major contributing variable, maximum radiation dose, Cox proportional hazards showed a consistent significant difference in time to pain relief favoring earlier SRS within 4 years of initial diagnosis (HR = 1.24 favoring ≤ 4 years, p = 0.033, adjusted) (Supplemental Table B).

Time to pain relief (in days since SRS) shown on product-limit failure Kaplan-Meier curves for the overall cohort (A), with 95% CIs shown, and stratified by patients who received SRS within 4 years of their initial diagnosis (blue) or beyond 4 years of their initial diagnosis (red, B). Hashmarks indicate censored values where follow-up ended. Figure is available in color online only.

Discussion

This large, multicenter, retrospective study successfully demonstrates the utility of early SRS in type I TN, in particular within 4 years of initial TN diagnosis. Prior to this study, two key publications had similarly aimed to define how timing influences response to radiosurgery in TN; the first, by Mousavi et al., demonstrated that type I TN patients who received SRS within 3 years of pain onset likewise experienced a shorter interval to pain relief and longer duration of pain control. ⁸ However, while only 42% of patients in the Mousavi et al. study were referred for SRS after becoming refractory to medical management, in our study, those refractory to medications comprised the vast majority of the cohort (81%). This difference in patient population likely contributes to the overall earlier (within 3 vs 4 years of pain onset) and faster (within 7 vs 21 days of SRS) pain response to SRS seen in the Mousavi et al. study; the progression of TN over time, especially when it is becoming increasingly refractory to medical management, is related to the development of demyelination and gliosis within the nerve, which likely also affects its response to radiosurgery. ¹⁰ Interestingly, a second landmark study, Lee et al. looked at the response to SRS in a cohort of TN patients who, at baseline, were all already medically refractory to medications or had failed prior surgical intervention. ⁹ In this study, the time to SRS from diagnosis that predicted faster pain relief was even longer than what was reported in our study (within 5 vs 4 years of pain onset), which is consistent with the thought that these patients might be more pathologically resistant to radiosurgery in their initial symptomatic years. Additionally, in the Lee et al. study, a pain history of ≤ 5 years was a significant predictor of durable pain relief, consistent with our findings. Nevertheless, this cohort is fundamentally different than that in the Mousavi et al. study and in our study patients in whom prior surgical therapy failed were included. Prior surgery is known to contribute to scar tissue formation, fibrosis, and subsequent nerve effect, which all likely influence a patient’s response to SRS.

The reported median maximum radiation doses in the aforementioned studies were 80 Gy in Mousavi et al. and 90 Gy in Lee et al., and the rates of new-onset facial numbness were 15% and 55%, respectively. The latter reported rate is relatively high compared to a reported median rate of new sensory symptoms of 19% in a systematic review across all TN subtypes. ¹¹ In our study, we reported a 24% new sensory symptom rate. Although a higher maximum radiation dose is known to correlate with improved and more durable pain control, as seen in our multivariable analysis as well, it is also associated with a higher incidence of new sensory symptoms. ^12,13 Interestingly, new-onset facial numbness following radiosurgery has been suggested to actually be predictive of a successful pain response and the durability of pain relief. ^14,15

Another critical goal of applying SRS to TN patients is to reduce pain medication dependence. It is well known that rates of pain medication failure in TN increase with time due to both tolerance and the increasing severity of the disease. ⁵ Additionally, adverse effects caused by these medications are common, particularly since tolerance stipulates that the dose required to maintain adequate pain control increases over time. Our study is the first to describe the influence of SRS on pain medication utilization of TN patients. In our study, the most commonly used medications prior to SRS were carbamazepine and GABAergics (gabapentin and pregabalin). Compared to pre-SRS baseline, the dependence on these medications reduced nearly by half across all patients (i.e., carbamazepine usage reduced from 52.2% to 27.5%, GABAergic usage reduced from 50.0% to 29.0%). This trend of highly significant reductions in usage was consistent both in the aggregate (mean number of medications pre-SRS vs last follow-up: 1.98 to 0.90, p < 0.0001) and across all pain medication subcategories (see Table 2). A critical focus of surgical intervention in TN should be the reduction of the patient’s reliance on pain medications and, in doing so, the reduction of the medication side effects experienced. According to our data, SRS is of value to achieving this goal.

While compelling, this study has several limitations that must be considered. First and foremost, its retrospective nature introduces inherent selection bias. Additionally, in an effort to control for a specific patient population (type I TN patients, in this case) to isolate treatment effect, these findings cannot be generalized to those with atypical TN, multiple sclerosis, or tumor-related TN. In atypical TN and TN related to multiple sclerosis, the response rate to SRS has been reported to be low. ^15,16 The relationship between tumor-related TN and pain relief is still being study. Inconsistent pain outcomes have been reported in patients who undergo SRS targeting the tumor alone; there is a small body of literature that advocates for the simultaneous treatment of tumor and nerve with SRS for more effective pain relief, but further study in this space is warranted. ^17–19 Lastly, the mean follow-up period in this cohort was 32 months, which is on par with previously series (i.e., 17 months, ⁹ 36 months ⁸ ), but it is limited in clarifying the true long-term durability of SRS for pain relief in these cases beyond a few years. Further long-term follow-up regarding SRS and the durability of pain relief would be particularly beneficial for this type I TN population whose patients tend to be older (mean age 70 years in this series) and are already otherwise experiencing polypharmacy. ²⁰

Our findings support the use of SRS earlier in the treatment of typical TN, especially within 4 years of initial pain onset and diagnosis. Treatment outcomes and the durability of pain response to SRS are related to the maximum radiation dose used. Additionally, the maximum dose was not a predictor of developing new sensory symptoms after SRS (p = 0.075), indicating that the median 80-Gy dose used in this cohort can be applied for durable pain relief and low risk. Our study is also the first to provide evidence that SRS reduces pain medication dependence in type I TN and thereby a reduction in medication-related side effects. We therefore advocate for earlier consideration of radiosurgery in the treatment course for typical TN and, to further verify these findings, call for additional research both in the context of a higher level of evidence (i.e., a randomized trial) and with longer-term follow-up.

Conclusions

In summary, the evidence presented in this study supports using SRS in the early management of patients with typical TN. Patients who underwent SRS within 4 years of diagnosis experienced a shorter interval to pain relief, with low risk of adverse sensory events, and were more likely to experience overall pain relief at their last follow-up. SRS also yielded significant decreases in adjunct medication utilization. We encourage physicians caring for patients with type I TN to refer for neurosurgical evaluation sooner rather than increasing medications to doses that lead to debilitating side effects.

Disclosures

Dr. Lunsford reports being a consultant for Insightec (DSMB); he owns stock in Elekta AB. Dr. Kondziolka reports receiving support for non–study-related research or clinical efforts that he oversees from Brainlab.

Author Contributions

Conception and design: Kondziolka, Mureb, Benjamin. Acquisition of data: Mureb, Strickland, Zada, Chang, Urgošík, Liščák, Warnick, Speckter, Eastman, Kaufmann, Patel, Feliciano, Carbini, Mathieu, Leduc, Nagel, Hori, Hung, Ogino, Faramand, Kano, Lunsford, Sheehan. Analysis and interpretation of data: Kondziolka, Mureb, Golub, Benjamin. Drafting the article: Kondziolka, Mureb, Golub. Critically revising the article: Kondziolka, Mureb, Golub, Benjamin, Gurewitz, Urgošík, Feliciano. Reviewed submitted version of manuscript: Kondziolka, Mureb, Golub, Benjamin, Gurewitz, Strickland, Zada, Chang, Liščák, Warnick, Speckter, Eastman, Kaufmann, Patel, Carbini, Mathieu, Leduc, Nagel, Hori, Hung, Ogino, Faramand, Kano, Lunsford, Sheehan. Statistical analysis: Kondziolka, Golub. Study supervision: Kondziolka, Mureb, Benjamin.

Supplemental Information