Abstract
Introduction
There is a lack of high-quality prospective, systematic studies using independent assessors of outcome of microvascular decompression as treatment for trigeminal neuralgia.
Methods
Clinical characteristics and outcome data were recorded prospectively from consecutive classical trigeminal neuralgia patients, using standardized interviews. Degree of neurovascular contact was evaluated by a 3.0 Tesla MRI blinded to symptomatic side. Patients were assessed before and 12 months after surgery by a neurologist.
Results
Twenty-six men and 33 women completed 12 months follow-up. Forty-one patients (69%) had an excellent outcome (no pain, no medication). Ten (18%) patients had a good outcome. Eight (12%) patients had no improvement or had worsening of pain. MRI showed neurovascular contact with morphological changes in 34 patients (58%). Odds ratio between neurovascular contact with morphological changes and excellent outcome was 4.4 (Cl 1.16–16.26), p = 0.029. Odds ratio between male sex and excellent outcome was 11.38 (Cl 2.12–59.52), p = 0.004. No significant association was found between excellent outcome and concomitant persistent pain, current age or disease duration.
Conclusion
Neurovascular contact with morphological changes and male sex are positive predictive factors for outcome of microvascular decompression. The findings enable clinicians to better inform patients before surgery.
Introduction
Trigeminal neuralgia (TN) is a painful neurological disease characterized by severe short-lasting stabbing unilateral facial pain. The majority of TN patients are treated medically, but approximately 30% of the patients are considered refractory to pharmacological treatment because of insufficient pain relief or side effects (1,2). Microvascular decompression (MVD) is the first-choice neurosurgical treatment option for medically refractory patients with an MRI verified neurovascular contact (NVC) (1).
During the last decades of evolving neuroimaging, pre-surgical MRI of the brain, brainstem, cranial nerves and cerebral vessels has become an essential part of the work-up (3). This development has led to an increasing understanding of the etiology of the disease. Due to this increased knowledge combined with previous studies on cadavers, it has become widely accepted that neurovascular contact (NVC) is the main but not the only etiological factor of TN (4,5).
The purpose of MVD is to alleviate the contact between the trigeminal nerve and the offending blood vessel. The mechanism of pain relief caused by the MVD is not fully understood, but it is hypothesized that alleviation of the offending blood vessel leads to remyelination of the damaged axons (5–7). Neuroinflammation and ischemia may also play a part in the pathophysiological mechanism and are thought to be reduced when the compression is alleviated (8). MVD is a highly effective treatment (9–11), but is an invasive procedure in the posterior fossa, and only approximately 70% have an excellent result when evaluated after 5–10 years (11–13). Identification of prognostic factors is therefore very important.
TN is more prevalent in women, with a male: female ratio of 1:1.5 (14), and a few studies have reported that the male sex is a positive prognostic factor for outcome of MVD (12,15,16). The reasons for this possible sex difference are unknown. The effects of severity of NVC on the prognosis of neurosurgical outcome have never been described.
The knowledge about MVD in classical trigeminal neuralgia is growing. Besides sex as a prognosticator for outcome, recent studies have identified age, duration of the disease, presence of the NVC and concomitant pain as predictors of the outcome of MVD (10,16–19). However, the majority of previous studies of MVD and prognostic factors have methodological limitations (20,21). They are often retrospective, and the use of international diagnostic guidelines and the method of neuroimaging vary. Most importantly, recording and evaluation of treatment outcome has not been done by independent assessors; that is, most often the outcome was evaluated by the operating neurosurgeon.
The objective of this paper was to describe prognostic factors for outcome of MVD in TN based on a consecutive and prospective series of TN patients evaluated by independent assessors.
Methods
Definition of the cohort
We consecutively evaluated patients with the diagnosis of TN seen at the Danish Headache Center (DHC), a tertiary medical referral centre for headache and facial pain. Patients who were treated with MVD at the Department of Neurosurgery between May 2012 and February 2016 were included in the study. The number of patients in the inclusion period determined the sample size. The study was observational, and patients were offered standard clinical care and laboratory workups according to current American Academy of Neurology and European Federation of Neurological Societies treatment guidelines (1).
Inclusion criteria were: a) diagnosis verified prior to surgery by a neurologist at the DHC, b) at least one pre-surgical out-patient evaluation of the clinical characteristics, pain level and previous treatments related to their TN, c) 3.0 Tesla MRI according to pre-defined protocol for the trigeminal nerve (22), d) at least 12 months follow up, and e) informed consent. Exclusion criteria were: a) history of previous MVD b) communication barriers c) unwillingness, or contraindications for open skull surgery.
The diagnosis was based on the International Classification of Headache Disorders ICHD-2 and ICHD-3 beta (23,24). As subtle sensory abnormalities are common in otherwise classic TN (14,25,26), we allowed for sensory abnormalities detected at clinical neurological examination, as previously described (26,27).
Pre-surgical neurological assessment
A standardized purpose-built semi-structured interview (14) was conducted by neurologists from DHC who paid special attention to onset, localization, character, intensity and frequency of pain and accompanying autonomic symptoms, to ensure correct diagnosis. Patients were systematically questioned about previous and current treatments, including effect, duration, dosage and side-effects of medical treatment and previous neurosurgical intervention(s). The clinical examination included full neurological examination with special emphasis on the sensory function of the trigeminal nerve, as previously described (14).
Pre-surgical neuroimaging
Patients underwent a 3.0 Tesla MRI scan according to a pre-defined protocol for the trigeminal nerve before referral to the Department of Neurosurgery.
As described previously in detail (22,27) the MRI protocol used for this study included axial thin- section T2-weighted GRASE (gradient and spin-echo) sequence encompassing the brainstem and the proximal and posterior fossa, sagittal T2-weighted turbo-spin-echo sequence covering the whole brain, 3D high spatial resolution heavily T2-weighted sequence (3D balanced fast field echo, BFFE) and 3D time of flight magnetic resonance angiography (s3DI MC HR) (22). Multiplanar re-constructions of the axial, sagittal and coronal plane were performed on the cisternal part of the NVC to ensure visualization of all planes and full length of the trigeminal nerve. In short, the following information was recorded from the scans: Degree, localization and type of NVC. An NVC was defined as contact between the trigeminal nerve and a blood vessel, without visible cerebrospinal fluid between the two structures. Degree of contact was graded as simple contact, displacement (displacement or distortion), or atrophy (reduced volume of the trigeminal nerve). NVC with morphological changes was defined as a contact with displacement and/or atrophy. It was described whether the contact was caused by an artery, vein or both. The imager was a 3.0 Tesla Phillips Achieva imager (Phillips Medical Systems) equipped with a 32-channel head coil with Multi Transmit parallel RF transmission. All scans were evaluated by the same experienced neuroradiologist (FW), who was blinded to symptom side. The observations of both the left and the right trigeminal nerves and compressing vessel(s) were registered in a standardized evaluation scheme.
Neurosurgical technique
The MVD was performed via a retrosigmoid craniectomy with a supracerebellar infratentorial approach. No retractors were used. The cerebellopontine angle was approached, and the trigeminal nerve and the compressing vessel(s) identified. In case of an arterial compression, the nerve was alleviated by transposition of the vessel. The vessel was transposed upwards and fixed to the tentorium with glue and a piece of Teflon. If transposition was not possible, a piece of Teflon was interposed between the nerve and the vessel. If the compression was caused by a vein, it was either transposed and fixed with glue or coagulated and divided to avoid avulsion.
Intraoperative findings with respect to localization, degree and type of NVC and possible periprocedural complications were registered in a standardized evaluation scheme, filled out by the neurosurgeon.
The procedure was performed by three highly trained neurosurgeons (JB, PR and JS) from the Department of Neurosurgery.
Follow-up and outcome assessments
All patients were assessed 3, 6 and 12 months after the MVD either by out-patient visits or telephone interviews. At each assessment, a neurologist or highly trained fellow in neurology at the DHC registered current pain according to the Verbal Numeric Rating Scale (VNRS) and the Barrow Neurological Institute (BNI) pain scale, complications to surgery, patient satisfaction and type and dose of medical treatment, if any. Twelve months after the MVD, a self-completed post-surgical questionnaire was mailed to the patients, including 31 quantitative and qualitative questions concerning a) current pain intensity, according to VNRS and BNI, b) complications, and c) patient satisfaction. The questionnaire was developed by Zakrzewska et al. (28) and was translated and modified by the authors of this paper. If there was a disagreement between the follow-up assessment and the patient questionnaire, data from the follow up scheme was reported in this study.
Seven-scale modified Barrow Neurological Institute pain intensity score.
The scale was modified by the authors of this paper from the original four-grade BNI. The modification was to ensure a more refined scale with a better distinction between the different outcomes of the procedure, including the possible outcome “Partial pain relief: No pain but daily medication required”, which is not included in the original four-scale BNI. Outcome of microvascular decompression one year after surgery is shown in the third column. Values represent numbers of patients (%); n = 59 trigeminal neuralgia patients. *The MVD was also regarded as failed if the patient had a recurrence of pain that needed re-operation, before their 12 months follow-up, regardless of type of re-operation.
BNI: Barrow Neurological Institute pain score.
Statistical analysis
Continuous and ranked data are summarized by descriptive statistics. Categorical variables are presented with frequency distributions (no, %) and with 95% confidence limits (Cl). A multiple logistic regression analysis with backward stepwise elimination was used to test for an association between a subset of predefined clinical variables and excellent outcome (no pain and no medication) of MVD. If there was a significant association (p < 0.05), the variables were retained in the model and variables with no significant association were excluded sequentially. The predefined independent variables were sex (male vs. female), age at operation (< 70 years vs. > 70 years), disease duration (< 2 years vs. > 2 years), degree of NVC at MRI (with vs. without morphological changes), and TN with purely paroxysmal pain vs. TN with concomitant persistent pain. As 27 patients are part of the same population as previously reported by our group, and as these studies reported that NVC with morphological changes is significantly more prevalent in males than in females (27), the multiple logistic regression model was controlled for this possible confounder by computing sex*neurovascular contact with morphological changes as an independent variable in the regression model. To test for associations between excellent outcome and the following variables: Autonomic symptoms, sensory deficits at pre-surgical examination, pure venous compression at MRI, and NVC with morphological changes found intraoperatively, we used the 2 × 2 contingency chi square test.
The chi-square test and Fisher’s test were used to assess post hoc explanatory analysis of differences in outcome between men and women. To estimate women’s menopausal state, the age of 50 years at the time of debut of TN was used as cut-off.
If more than 10 % of the clinical data from the interview, follow-up schemes or the questionnaire were missing, the patient was excluded from the analyses. Missing data was considered missing at random. p-values are reported as two-tailed with a level of significance of 5%. Analyses were carried out using SAS 9.4 (SAS Institute Inc., NC, USA). We used STROBE reporting guidelines (31) as the checklist for the reporting of our findings.
Ethics and Institutional Review Board approval
The Danish National Committee on Health Research Ethics, project ID nr. H-15008830, confirmed that the project did not need ethical approval, as it was observational and based on routine clinical care and laboratory workup. All patients gave their written informed consent before their data were analyzed, according to the regulations of the Danish Data Protection Agency.
Results
From May 2012 to February 2016, 4,439 patients were referred to DHC. Of these, 359 patients were diagnosed with TN. The majority of patients, 264 (74%), continued medical treatment only, while 95 (26%) were referred to neurosurgery. Of these, 72 (76%) patients were treated with MVD, of whom five (7%) patients had a history of previous MVD. Two (2%) patients were treated with glycerol blockade and 21 (22%) had a balloon compression. Thirteen (18%) patients were excluded from the present study. A total of 59 TN patients, 33 women and 26 men, were included in the study. The inclusion of patients is outlined in Figure 1. Post-hoc testing of the diagnosis of the patients who were diagnosed according to ICHD-2 criteria showed that they also met the ICHD-3 beta criteria. Demographics, baseline clinical characteristics and comorbidities are shown in Table 2.
Flowchart of inclusion of trigeminal neuralgia patients. Demographics, baseline clinical characteristics and comorbidities. Values represent numbers of patients (%) unless otherwise indicated. n = 59 trigeminal neuralgia patients. Calculations on concomitant persistent pain and sensory abnormalities only include patients with no history of previous trigeminal neuralgia surgery, n = 56. History of previous radiofrequency thermocoagulation n = 1, history of previous glycerol injection, n = 2, before microvascular decompression. Cl = confidence limits. TN = trigeminal neuralgia. Three patients had not tried sodium channel blockers at the time of their first outpatient visit. Chi square, bFisher’s exact test.
Prognostic factors for microvascular decompression
One year after surgery, the outcome was reported as excellent (BNI I) in 41 patients (69%). Fifty-one patients (87%) reported a clinically relevant pain reduction, equal to excellent or good outcome (BNI I–IIIB). In one (2%) patient, the outcome was poor and in seven (12%) patients, the outcome was considered as failure either due to aggravation of pain (two patients (3%)), or because they had a subsequent neurosurgical procedure before the 12 months follow-up (five patients (9%)) (Table 1). Four of the five patients who had new neurosurgical treatment were women.
Associations between clinical characteristics and excellent outcome.
n = 59 trigeminal neuralgia patients, of whom 41 had an excellent outcome.
The analysis of the association between the prognostic variable and excellent outcome was done by multiple logistic regression with backwards elimination. p-values are reported as two-tailed with a level of significance of 5%. OR: odds ratio; Cl: confidence limits.
Good outcome (BNI II–IIIB) was reported by 10 patients, of whom nine were women. Post hoc multiple logistic regression analysis on the association between “excellent – good outcome (BNI I–IIIB)” showed a tendency towards an association between male sex and good outcome (OR 6.7, CL 0.77–58.71, p-value 0.08). All the same clinical variables were included in the analysis model. Severe NVC was non-significant in this post hoc analysis.
In 34 patients, the MRI showed NVC with morphological changes (Figure 2), of which 28 patients (82%) had an excellent outcome. In contrast, only 13 (52%) of the 25 patients with simple NVC had an excellent outcome. Of the 26 men included in the study, 24 (92%) had an excellent outcome. In contrast, 33 women were included, of whom 17 (51%) had an excellent outcome.
Pre-operative MRI of a patient with left-sided trigeminal neuralgia.
NVC with morphological changes was found in 17 men, of whom 16 (94%) had an excellent outcome. NVC without morphological changes was found in nine men, of whom eight (89%) had an excellent outcome. In contrast, 17 women had an NVC with morphological changes, of whom 12 (71%) had an excellent outcome; NVC without morphological changes was found in 16 women, of whom five women (31%) had an excellent outcome (Figure 3).
Sex, neurovascular contact and outcome of microvascular decompression.
There was no significant association between purely paroxysmal pain (yes vs. no) and excellent outcome, between age (below vs. over 70 years) and excellent outcome, or between disease duration (below vs. above 2 years’ duration) and excellent outcome (Table 3). There was no significant association between NVC with morphological changes found intraoperatively (yes vs. no) and excellent outcome (OR 0.78 (Cl 0.21-2.92), p = 0.71) (). The 2 × 2 contingency chi square tests showed no significant association between autonomic symptoms, sensory deficits at pre-surgical examination, pure venous compression and excellent outcome.
Post hoc exploratory chi-square tests did not show any significant association between women above 50 years of age at the time of the debut of their TN and excellent outcome, compared to women below 50 years of age (Chi-square value = 1.5, p = 0.22).
Post hoc correlation analysis showed no significant correlation between age and outcome according to BNI I–VB (correlation coefficient r = −0.19, p = 0.14). There was no significant correlation between disease duration and outcome according to BNI I-VB (correlation coefficient r = −0.07, p = 0.55).
Pre-surgical neuroimaging findings
Pre-surgical neuroimaging findings.
Values represent numbers of patients (%) unless otherwise indicated. n = 59 trigeminal neuralgia patients.
*In one patient, the protocolled 3.0 Tesla MRI did not show any neurovascular contact. Neurovascular contact without morphological changes was defined as contact between the trigeminal nerve and a blood vessel, without visible cerebrospinal fluid between the two structures. Neurovascular contact with morphological changes was defined as a contact with displacement and/or atrophy. Root entry zone was defined as from the site of entry of the trigeminal nerve into the pons to maximum 7 mm peripherally along the nerve. Peripheral contact was defined as > 7 mm from the site of entry of the trigeminal nerve into the pons.
**Two patients (3%) had a neurovascular contact in the peripheral segment of the nerve and three patients (5%) had neurovascular contact both in the root entry zone and in the peripheral segment.
Neurosurgical findings
The neurosurgeons reported that all 59 patients had an NVC on the operated side. NVC caused by a single artery was the most frequently reported, in 27 (46%). The most frequently reported artery was the superior cerebellar artery in 45 (76%), anterior inferior cerebellar artery caused NVC in five (8%) patients, the trigeminal artery in one (2%), and the vertebral artery in two patients (3%). The artery could not be specified in five patients (14%).
Discussion
This prospective study using blind state-of-the-art neuroimaging and neurologists for evaluation of outcome after MVD demonstrated that NVC with morphological changes and male sex are highly predictive of excellent outcome of MVD. Concomitant persistent pain, age, and duration of disease did not influence the operative results. Supporting previous literature, we demonstrate that MVD is generally an effective neurosurgical treatment option for medically refractory TN when evaluated 1 year after surgery.
Our results enable the clinicians to differentiate TN patients before surgery, identifying those who are expected to have an excellent outcome. International pain experts have asked for this information for years (1,20). Moreover, our results provide patients and clinicians with a better basis for decisions about MVD.
Male sex and neurovascular contact with morphological changes predicts excellent outcome
Diffusion Tensor Imaging studies show that pathological changes of the trigeminal nerve before MVD normalize after MVD (32). This was interpreted as a sign of remyelination and reduced neuroedema at the site of NVC (32). This finding could explain why patients with severe NVC have a better outcome according to the present and previous studies (12,33).
Our findings also show a more complex relationship between NVC and TN than previously found. NVC with morphological changes is central to TN etiology, but the male/female sex seems to be equally important. Even when controlled for degree of NVC, the male sex was associated with a considerably higher odds ratio for an excellent outcome. Men without morphological changes of the trigeminal nerve also had a higher rate of an excellent outcome, and even compared to women with morphological changes of the trigeminal nerve, the outcome was better. One may hypothesize that men may have a more mono-factorial etiology (NVC), than women in whom other factors are equally important.
Gender differences in pain and pain-relief has been described and debated for decades (34,35). There is no current consensus on whether sex is a prognostic factor for MVD as treatment for TN. It remains speculative whether the considerable difference in outcome in men and women in the present study could reflect sex differences in hormonal status or genetics or is due to differences in patients’ emotional state (36,37). Our post hoc explanatory analysis did not show any association between women above 50 years of age (post-menopausal) and excellent outcome. However future studies with larger cohorts are needed to further explore these findings.
Individualized information as part of joint decision making
In a recently published paper, Cruccu et al. suggested subdividing TN into a classical form, with NVC with morphological changes of the trigeminal nerve, and an idiopathic form without morphological changes of the nerve (38). They assumed that MRI findings are crucial for disease mechanisms and treatment results. We support the notion that morphological changes of the trigeminal nerve are an important and useful objective finding when planning treatment for this patient group. However, the suggested subgroups of “classical” and “idiopathic” TN based on degree of NVC (38) are not in agreement with our results.
The current results indicate that the dichotomization into classical and idiopathic TN may be too simple and therefore should be modified; in this study, some NVCs without morphological changes appear to be important to disease etiology (especially in men) and other NVCs without morphological changes appear to be irrelevant to disease etiology (truly idiopathic TN, seen especially in women). While it is important not to perform MVD in TN patients where the NVC is not a part of the etiology but a variation of normal anatomy, it is also important not to exclude patients with NVC without morphological changes from being considered for MVD. However, currently it is not possible to demonstrate whether an NVC without morphological changes is truly without relevance for the individual patient. Further development of imaging techniques, for example diffuse tensor imaging, may enable us to do this in the future.
Excellent response is also frequent in patients with concomitant pain
It has been debated whether only patients with purely paroxysmal pain should be offered MVD, as presence of concomitant persistent pain (also described as atypical trigeminal neuralgia or trigeminal neuralgia type 2) has been reported as a negative prognostic factor of outcome (39,40). Our data demonstrate that patients with concomitant persistent pain have as good results of MVD as patients with purely paroxysmal pain. It has been hypothesized that concomitant pain is a sign of chronification due to increasing mechanical compression of the trigeminal nerve (41). However, our results, in line with other recent research, indicate that concomitant pain is a result of central sensitization rather than a sign of late chronification of the disease (27,42). An explanation for the discrepancy in results between the current and previous papers is that the current paper used logistic regression to control for confounders. Sex and degree of NVC was maintained in all analyses, as they were considered a priori to be important. In a previous study (43) we found that women had a higher prevalence of concomitant persistent pain. Thus, findings from other previous neurosurgical studies might actually reflect that more women included in the studies had concomitant persistent pain and therefore less optimal outcome. Yet, the current results reflect outcome of MVD in a relatively small cohort. Thus, these findings might change if similar studies are done in a larger cohort.
Strengths and weaknesses
Our study was a prospective cohort study as recommended to address prognostic questions (44). The outcome evaluated by independent assessors and the systematic blind assessments of MRI reduced potential bias markedly and are the major strengths of this study.
In this study, we exclusively report the excellent outcome 1 year after MVD. However, previous studies report the recurrence rate as approximately 10% in 5 years (12,16). Therefore, our results are early outcome results only, which is a limitation. On the other hand, even if pain recurs later, one year of pain freedom is clinically very important for the individual patient. Multicenter long-term prospective follow-up studies using independent assessors of outcome and a larger volume of patients are greatly needed as outcome may change in the following 2–10 years after surgery. This is particularly necessary to evaluate complications that fortunately occur rarely. Because of the too-small numbers we do not report complications in this paper and the results only reflect the effect regarding reduced TN pain.
Clinical implications
Men have considerably better chance of excellent outcome of microvascular decompression, than women. Patients with morphological changes of the trigeminal nerve have a better chance of excellent outcome of microvascular decompression than patients without morphological changes. Excellent outcome of MVD is seen also in patients with concomitant persistent pain, advanced age and long duration of pain. Study results enable clinicians to stratify trigeminal neuralgia patients before surgery, identifying those who are expected to have an excellent outcome.
Footnotes
Acknowledgements
The authors thank Jacob Bertram Springborg (JS) for his dedication to this project, for undertaking the MVD of TN patients and his registration of perioperative findings. We also thank the medical secretaries Ane Lundgaard Dahl and Dianna Bartolin for their valuable help with the data extraction.
Authorship
All authors of the paper meet all criteria set by the International Committee of Medical Journal Editors (ICMJE) authorship guidelines.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The patient organization Trigeminus Foreningen provided financial support for the study. The funding sources had no role in the study. The authors received no other financial support for the research, authorship, and/or publication of this article.
Trial Registration
The current study was observational, and patients were offered standard clinical care and laboratory workups according to current American Academy of Neurology and European Federation of Neurological Societies treatment guidelines. As the study was observational and initiated in 2012, we did not register the study at a trial site (e.g. ClinicalTrials.gov).