Pain Outcomes Following Microvascular Decompression for Drug-Resistant Trigeminal Neuralgia: A Systematic Review and Meta-Analysis

Abstract

BACKGROUND

Microvascular decompression (MVD) is a potentially curative surgery for drug-resistant trigeminal neuralgia (TN). Predictors of pain freedom after MVD are not fully understood.

OBJECTIVE

To describe rates and predictors for pain freedom following MVD.

METHODS

Using preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines, PubMed, Cochrane Library, and Scopus were queried for primary studies examining pain outcomes after MVD for TN published between 1988 and March 2018. Potential biases were assessed for included studies. Pain freedom (ie, Barrow Neurological Institute score of 1) at last follow-up was the primary outcome measure. Variables associated with pain freedom on preliminary analysis underwent formal meta-analysis. Odds ratios (OR) and 95% confidence intervals (CI) were calculated for possible predictors.

RESULTS

Outcome data were analyzed for 3897 patients from 46 studies (7 prospective, 39 retrospective). Overall, 76.0% of patients achieved pain freedom after MVD with a mean follow-up of 1.7 ± 1.3 (standard deviation) yr. Predictors of pain freedom on meta-analysis using random effects models included (1) disease duration ≤5 yr (OR = 2.06, 95% CI = 1.08-3.95); (2) arterial compression over venous or other (OR = 3.35, 95% CI = 1.91-5.88); (3) superior cerebellar artery involvement (OR = 2.02, 95% CI = 1.02-4.03), and (4) type 1 Burchiel classification (OR = 2.49, 95% CI = 1.32-4.67).

CONCLUSION

Approximately three-quarters of patients with drug-resistant TN achieve pain freedom after MVD. Shorter disease duration, arterial compression, and type 1 Burchiel classification may predict more favorable outcome. These results may improve patient selection and provider expectations.

Graphical Abstract

Graphical Abstract.

ABBREVIATIONS

BNI

barrow neurological institute

CI

confidence interval

CSF

cerebrospinal fluid

PRISMA

preferred reporting items for systematic reviews and meta-analyses

MRI

magnetic resonance imaging

MVD

microvascular decompression

OR

odds ratios

SCA

superior cerebellar artery; TN, trigeminal neuralgia

Trigeminal neuralgia (TN) is characterized by paroxysmal, lancinating facial pain along the distribution of one or multiple divisions of the trigeminal nerve.¹ It is most often caused by compression of the trigeminal nerve by a blood vessel in the prepontine cistern.² Pain episodes are generally described as debilitating, severe, and sharp while typically lasting only seconds to minutes.¹ When lancinating facial pain predominates, it is described as typical facial pain, or “Burchiel type 1” classification,³ and can be triggered by activities such as teeth brushing, eating, or talking. Conversely, atypical TN pain is predominantly described as a constant burning pain accompanying lancinating pain and is deemed “Burchiel type 2” classification.^1,3

Medical treatment generally consists of carbamazepine or other antiepileptic drugs such as gabapentin, pregabalin, or oxcarbazepine.¹ Response to carbamazepine is typically favorable, with some studies reporting ∼90% of patients experiencing a reduction in number of episodes, but the therapeutic window is relatively small and side effects are common.⁴ In a meta-analysis of medical management in 639 TN patients, 44% reported adverse effects to the drug, which consisted primarily of vertigo, somnolence, nausea, and fatigue.⁴ Refractory cases or those with intolerable medication side effects should be referred for neurosurgical evaluation.

Microvascular decompression (MVD) is the most frequently used treatment for drug-resistant TN.⁵ Microsurgical exploration leads to identification of the offending vessel (eg, artery or vein) and isolation of the nerve using Teflon padding or other agents.⁶ Long-term pain outcomes are generally favorable, but there is significant variability in outcome rates reported throughout the literature, likely due to patient and center variability. For instance, whereas some studies report pain free rates as high as 90 to 93% after surgery,^7-9 others find that only 50 to 71% of patients achieve pain freedom.^10,11 Furthermore, whereas a few systematic reviews have evaluated subtopics related to MVD for TN,^12-19 none have quantitatively evaluated predictors of favorable outcome. Therefore, overall rates and predictors of pain freedom in MVD for TN remain incompletely understood.

The objective of this study is to examine rate and predictors of pain freedom in drug-resistant TN patients who undergo MVD. A number of variables are collected and examined, including disease duration, type of nerve compression, type of facial pain, and the number of nerve root branches involved. Common complications associated with MVD are also briefly reviewed. To our knowledge, this is the first large-scale formal meta-analysis evaluating both outcome rates and predictors in MVD for TN.

METHODS

Literature Review

PubMed, Cochrane Library, and Scopus were searched for high-quality primary clinical studies describing pain outcomes following MVD for TN in March of 2018 (Figure 1). The search terms “microvascular decompression” and “trigeminal neuralgia” with the filters for “English language,” “human subjects,” and publication dates “1988-2018” were applied to all search engines. This returned 922 unique citations after exclusion of redundant studies. The inclusion criteria were the following: (1) primary clinical studies reporting pain outcomes following MVD for TN (ie, book chapters or reviews were not included); (2) the publication dates were later than 1988 to reduce publication year bias; (3) only adult human subjects were included; and (4) the articles were written in the English language. There were 225 studies reviewed by full text. Then the following exclusion criteria were applied: (1) a study had absent or unclear pain outcome data (ie, unable to dichotomize outcomes into pain free vs persistent pain); (2) the study had less than 10 patients, which might increase publication biases (eg, case reports were excluded); (3) patient data were redundant with another manuscript; and (4) the mean or median follow-up was shorter than 1 yr. Ultimately, 46 studies met all criteria and were used for analysis. PRISMA guidelines were followed for this systematic review.²⁰

FIGURE 1.

Flow chart illustrating the study selection process.

Data Collection

The primary outcome was pain outcome at the last follow-up, which was dichotomized into “pain free” vs “persistent pain.” Pain free was defined as a barrow neurological institute (BNI) score of 1. Patient data were examined for redundancy with another publication prior to collection. The following variables were collected if they were unambiguously associated with pain outcome: (1) age; (2) sex; (3) disease duration; (4) vessel causing compression; (5) TN type; (6) preoperative magnetic resonance imaging (MRI) results; (7) number of trigeminal nerve branches compressed; (8) superior cerebellar artery (SCA) involvement; and (9) the side of surgery. If disease duration data were unclear, or were dichotomized at a number other than 5 yr, these data were excluded from the disease duration calculation. The vessels causing compression were divided into artery, vein, mixed, or other. The type of TN was divided into Burchiel types 1 and 2, which included typical and atypical TN pain syndromes, respectively. The number of trigeminal nerve branches compressed was dichotomized into one vs multiple branches. When available, complication rates and types were also recorded for a subset of patients, though not enough data were available to perform a systematic review.

Possible bias was assessed and recorded for each included study's methodology. Interviewer bias was defined as “a systematic difference between how information [was] solicited, recorded or interpreted,” which occurred when the interviewer was not blinded.²¹ Possible recall bias occurred when the outcomes of treatment affected the patient's memory for previous events. Selection bias referred to patients dropping out or failing to follow-up or if the exposure and outcome had already occurred prior to patient selection.²¹ Response bias described the bias that occurred with self-report and surveys.²²

Statistical Analysis

Pain outcomes were stratified across each variable before preliminary analysis, which performed for summary purposes and guidance of meta-analysis. Student t-tests and chi-square tests were used to evaluate continuous and categorical data, respectively. Pearson correlation tests were utilized to evaluate the relationship between pain outcomes and sample size or publication year. Means and standard deviations were calculated. Factors that were potentially associated with pain outcomes on preliminary analysis (P < .05) were subjected to formal meta-analyses. Heterogeneity across studies was tested using Cochran's Q and I² tests. Mantel–Haenszel tests calculated combined odds ratios (OR) and 95% confidence intervals (CI). A random, rather than fixed effects, model was used because clinical heterogeneity among studies allowed us to accurately assume that there was likely a distribution of ORs rather than a single common one.²³ Funnel plots of ORs were assessed for asymmetry for potential biases. Wizard Pro 1.8.28 and Review Manager v5.3 (Nordic Cochrane Centre, Rigshospitalet, København, Denmark) were used for all statistical analyses.

RESULTS

There were 3897 patients included from 46 studies (Table 1).^{7-11,15,24-63} The overall pain freedom rate was 75.8% at a last follow-up of 1.7 ± 1.3 (mean ± standard deviation) yr. There was no significant correlation between the pain freedom rate and the year of publication [r(44) = 0.24; P = .12] or sample size [r(44) = 0.19; P = .20]. Preliminary analysis was done to guide formal meta-analysis, and a summary is shown in Table 2. For demographics, there were no significant relationships between age and gender with pain outcome (Table 2, demographic). For disease-related variable diseases, the duration in years, the type of vessel compressing the nerve, and the type of TN were all associated with varied pain outcomes (Table 2, disease). For the anatomic variables, the preoperative MRI findings, the number of TN branches compressed, and SCA involvement were also associated with pain outcomes, whereas the side of compression was not (Table 2, anatomy).

TABLE 1.

Studies With Sample Size, Level of Evidence, and Potential Biases

Study	n	Evidence level	Potential biases
Broggi et al49	36	3	Interviewer, recall, responder
Burchiel et al55	36	3	Interviewer, recall, responder, selection
Chai et al36	157	3	Interviewer, recall, responder
Chakravarthi et al7	39	3	Interviewer, recall, responder
Chen et al32	322	3	Interviewer, recall, responder
Cheng et al58	94	3	Recall, responder, selection
Dahle et al54	57	3	Interviewer, recall, responder
Duan et al29	26	3	Interviewer, recall, responder, selection
Dumot et al26	55	3	Interviewer, recall, responder
Ferroli et al43	401	3	Interviewer, recall, responder, selection
Gunther et al9	283	3	Recall, responder
Hai et al10	26	3	Interviewer, recall, responder
Hamlyn et al63	33	2	Interviewer, responder
Han et al46	10	3	Interviewer, recall, responder, selection
Han-Bing et al44	167	3	Interviewer, recall, responder
Hong et al40	13	3	Interviewer, recall, responder
Inoue et al25	31	3	Interviewer, recall, responder, selection
Kabatas et al15	10	3	Interviewer, recall, responder
Kato et al60	13	3	Interviewer, recall, responder
Leal et al34	50	2	Interviewer, recall, responder, selection
Li et al48	62	3	Interviewer, recall, responder
Li et al35	13	3	Interviewer, recall, responder
Li et al30	19	3	Interviewer, recall, responder, selection
Liao et al27	246	3	Interviewer, recall, responder, selection
Ma et al37	6	3	Interviewer, recall, responder
Masuoka et al31	50	3	Interviewer, recall, responder, selection
Mendoza et al52	134	3	Recall, responder
Miller et al45	95	3	Recall, responder
Obata et al57	51	3	Interviewer, recall, responder
Panczykowski et al28	22	3	Interviewer, recall, responder, selection
Ruiz-Juretschke et al59	125	3	Interviewer, recall, responder
Salama et al61	21	2	Interviewer, responder
Sandel et al8	226	3	Recall, responder
Sekula et al42	14	3	Recall, responder
Sekula et al41	89	2	Recall, responder
Simms et al39	64	3	Interviewer, recall, responder, selection
Sindou et al62	362	2	Interviewer, responder
Slettebo et al51	25	2	Responder
Sun et al53	56	3	Interviewer, recall, responder
Sun et al24	12	3	Interviewer, recall, responder
Tacconi et al50	8	3	Interviewer, recall, responder
Walchenbach et al11	58	3	Recall, responder
Yang et al38	10	3	Interviewer, recall, responder
Yang et al33	223	3	Interviewer, recall, responder, selection
Zacest et al47	13	3	Interviewer, recall, responder
Zhao et al56	34	3	Interviewer, recall, responder

Overall pain freedom rate was 75.8% at last follow-up of 1.7 ± 1.3 (mean ± SD) yr.

TABLE 2.

Preliminary Analysis of Potential Factors Associated With Pain Outcome

			Pain free	Persistent pain	P Value	n
Demographic	Age (yr)		57.5 ± 13.5	55.2 ± 13.6	.17	302
	Sex	Male	108 (71.1)	44 (29.9)	.17	420
		Female	196 (73.1)	72 (26.9)
Disease	Disease duration (yr)		4.6 ± 4.0	6.4 ± 5.5	<.01	289
	Compression	Artery	518 (77.5)	150 (22.5)	<.01	1031
		Vein	142 (64.0)	80 (36.0)
		Mixed	67 (57.8)	49 (42.2)
		Other	8 (32.0)	17 (68.0)
	Type of TN	One	614 (73.4)	223 (26.6)	<.01	1140
		Two	184 (60.7)	119 (39.3)
Anatomy	Preoperative imaging	Possible Compression	158 (78.6)	43 (21.4)	<.01	269
		Normal	40 (58.8)	28 (41.2)
	Number of branches compressed	Multiple	141 (68.5)	65 (31.5)	.04	397
		Single	148 (77.5)	43 (22.5)
	SCA involved	Yes	242 (78.3)	67 (21.7)	<.01	526
		No	148 (68.2)	69 (31.8)
	Side	Right	135 (70.7)	56 (29.3)	.57	316
		Left	92 (73.6)	33 (26.4)
Total			2953 (75.8)	944 (24.2)

SCA, superior cerebellar artery; TN, trigeminal neuralgia.

Data are mean ± SD for continuous variables N (%) for categorical variables. Bolded numbers are statistically significant.

All factors with potential association with pain freedom were subjected to formal meta-analysis (Figure 2). Disease duration was dichotomized into ≤5 yr vs longer, as the mean duration was 5.2 ± 4.6 yr for all included patients. Tests of heterogeneity were significant for vessel type causing nerve compression (χ² = 29.25; df = 17; I² = 42%; P = .03), preoperative MRI findings (χ² = 14.02; df = 2; I² = 86%; P < .01), and type of TN (χ² = 20.03; df = 7; I² = 65%; P < .01), but not significant for disease duration (χ² = 6.25; df = 10; I² = 0%; P = .79), the number of nerve branches compressed (χ² = 7.99; df = 12; I² = 0%; P = .79), and SCA involvement (χ² = 14.97; df = 11; I² = 27%; P = .18).

FIGURE 2.

Meta-analyses examining factors associated with pain freedom. Significant predictors of pain freedom were 5 yr or shorter disease duration A, artery involvement B, type 1 TN C, and SCA involvement F. A, Disease duration: 5 yr or shorter over longer than 5 yr disease duration. Heterogeneity: Tau2 = 0.00; Chi2 = 6.25, df = 10 (P = .79); I2 = 0%. Test for overall effect: Z = 2.18 (P = .03). B, Artery involvement over noninvolvement in compression. Heterogeneity: Tau2 = 0.74; Chi2 = 28.39, df = 14 (P = .01); I2 = 51%. Test for overall effect: Z = 3.43 (P = .0006). C, Type 1 over type 2 Burchiel TN classification. Heterogeneity: Tau2 = 0.22; Chi2 = 9.99, df = 6 (P = .12); I2 = 40%. Test for overall effect: Z = 3.82 (P = .0001). D, Suggested compression over no suggested compression on preoperative MRI. Heterogeneity: Tau2 = 4.10; Chi2 = 14.02, df = 2 (P = .0009); I2 = 86%. Test for overall effect: Z = 0.60 (P = .55). E, Single branch over multiple branch compression. Heterogeneity: Tau2 = 0.00; Chi2 = 7.83, df = 11 (P = .73); I2 = 0%. Test for overall effect: Z = 1.79 (P = .07). F, Superior cerebellar artery involvement over noninvolvement. Heterogeneity: Tau2 = 0.27; Chi2 = 11.58, df = 9 (P = .24); I2 = 22%. Test for overall effect: Z = 2.45 (P = .01).

Utilizing a random effects meta-analysis model for all variables, significant predictors of pain freedom were “5 years or shorter” over “longer than 5 years” disease duration (OR 2.06, 95% CI 1.08-3.95; P = .03), “artery involvement” over “noninvolvement” in nerve compression (OR 3.35, 95% CI 1.91-5.88; P < .01), “type 1” over “type 2” TN classification (OR 2.83, 95% CI 1.32-4.67; P < .01), and “SCA involvement” over “noninvolvement” (OR 2.02, 95% CI 1.02- 4.03; P = .04). Insignificant predictors were “compression suggested” over “normal” based on preoperative MRI (OR 2.19, 95% CI 0.17-28.14; P = .55) and “single” over “multiple” TN branch compression (OR 1.67, 95% CI 0.94-2.99; P = .08).

A systematic review and meta-analysis of potential complications was not performed, but complications from example studies are summarized in Table 3.

TABLE 3.

Examples of Major and Minor Complications Reported During Long-Term Follow-Up in 5 Studies

Study	Major complications	Minor complications
Yang et al33	Hearing loss (0.9%), partial facial nerve palsy (0.4%), meningitis (0.9%), CSF leak (0.9%)	Headaches (3.6%), vomiting (4.0%), facial dysesthesia (2.7%), ataxia (0.9%)
Sandel et al8	Cerebellar hematoma (1.2%), CSF leak (2.9%), abacterial meningitis (0.4%), facial paresis, death (0.4%)	Wound infection (1.6%), facial numbness (0.9%), dizziness (1.8%), hearing impairment (2.7%), tinnitus (0.4%)
Sekula et al41	Hyponatremia (5.6%)	Trigeminal numbness (13.9%), dysesthesias (5.7%)
Ferroli et al43	Hearing loss (1%), CSF fistula (4.2%), posterior fossa subdural hematoma (0.2%), bacterial meningitis (0.4%)	Trigeminal numbness (5.5%), transient diplopia (3.2%), mild hearing loss (5.2%)
Sindou et al62	Trochlear nerve palsy (0.83%), facial nerve palsy (0.83%), hearing loss (1.9%), death (0.44%)	Trigeminal numbness (3.04%)

CSF, cerebrospinal fluid.

DISCUSSION

In this systematic review and meta-analysis including 3897 patients from 46 studies, we described rates and predictors of pain freedom following MVD for drug-resistant TN. Overall, long-term pain freedom was favorable, with 76.0% of patients achieving pain freedom at last follow-up, which is similar to previous reviews.^12,13 To our knowledge, this is the first formal meta-analysis to examine predictors of outcome in MVD for TN, and we found that a disease duration of ≤5 yr, arterial involvement, SCA involvement, and type 1 facial pain classification were predictors of pain freedom.

Predictors of Pain Freedom

Disease duration ≤5 yr was predictive of postoperative pain freedom, which was consistent with the literature. Five years was chosen as the dichotomization point given mean disease duration of about 5 yr and for ease of future clinician use. In 2 previous studies, a disease duration of shorter than 7 or 8 yr was a predictor of favorable outcome after MVD, with the best outcomes seen with a duration shorter than 4 yr.^64,65 It was postulated that this might be due to prolonged root damage from neurovascular compression. Theoretically, demyelination after a certain disease duration may be severe enough to induce an irreversible central sensitization even after decompression.⁶⁶ The exact reason for persistent pain after technically successful MVD remains unknown and will require further research.

Our results showed that arterial rather than venous compression was predictive of pain freedom. There are various opinions about whether veins alone generate enough force to cause significant nerve compression or if they are an underrated or commonly missed cause of compression.^67,68 Conversely, the “ignition theory” states that even minimal compression from a vein at the nerve root is enough to produce the symptoms.² We found that arterial involvement predicted pain freedom after MVD and that SCA involvement specifically was a positive predictor. Findings in the literature are mixed. Some studies have shown that MVD alleviating venous compression has led to positive long-term pain outcomes.^25,26 Other studies report MVD of arterial compression was significantly better than venous compression,⁴⁴ and recurrence of pain was seen less frequency in arterial compression than in venous or mixed cases.⁶³ In a long-term study of 425 patients with refractory TN, arterial compression was a factor that predicted pain freedom following MVD.⁶⁹ No studies in our literature search reported the SCA individually as a factor in pain freedom after MVD. Overall, this finding is limited by physician interpretation of exactly which vessel is causing compression based on imaging and intraoperative findings, which may be difficult to discern.

We found that “type 1” pain was a predictor of pain freedom compared to “type 2” classification, which is consistent with the literature.⁸ In this cohort of patients, type 1 TN had a rate of pain freedom of 73.4% compared to 60.7% in type 2 TN. For example, one study reported that the rate of pain freedom in type 1 TN (73.7%) was twice as high as type 2 TN (34.7%).⁷⁰ In another prospective cohort study, patients with type 1 presentation were 3 times more likely to be pain free compared to type 2 presentation.⁷¹ The authors hypothesized that this finding may be due to the observation that type 1 disease was twice as likely to be associated with arterial compression (ie, a positive outcome predictor), whereas type 2 patients were more likely to have venous compression (ie, a negative outcome predictor). Furthermore, type 1 patients were more likely to have initial pain relief and lower recurrence rates at 2 yr.⁴⁵ Overall, MVD should be pursued more cautiously in the setting of atypical TN symptoms, and patients should be counseled that a favorable outcome may be less likely.

Complications

A summary of complications is described in Table 3. Minor complications, such as facial numbness, dysesthesia, and vertigo, were relatively common. Facial numbness was one of the most commonly reported minor complications in the literature as well; one large study of 401 patients followed over a 10-yr period reported facial numbness in 5.5% of patients,⁴³ whereas a separate prospective study of 89 patients observed it occurred in 13.9% of individuals.⁴¹ Other common minor complications in the literature include facial dysesthesia (5.3-5.7%),^41,43 minor hearing impairment (2.7%),⁸ and dizziness (1.8%).⁸ More serious complications, such as cranial nerve palsies, hearing loss, cerebrospinal fluid (CSF) leak, and infection, are rare in the literature, with rates typically lower than 3%.^12,33,43 Mortality is extremely rare, with larger studies reporting 0 to 0.4% mortality.^12,17,41 Overall, MVD is a safe surgery, with the most common complications being facial numbness, dysesthesia, and vertigo, and more serious complications, such as hearing loss, CSF leak, infection, and mortality are far less common.

Study Limitations

There are several limitations of this meta-analysis to consider. First, the majority of included studies were retrospective, which is consistent with the majority of neurosurgical literature.^12,13,15-19 A larger prospective study with controlled conditions would offer a higher quality of data to clarify rates and predictors of outcomes. Second, although we acquired data from 3897 patients, not every variable was reported for each patient, and thus, each analysis comprised only a subset of data. This limitation is also inherent to neurosurgical meta-analyses.⁷² Third, the heterogeneity of the data is a limitation inherent in meta-analysis. For instance, patients with a history of multiple sclerosis and postherpetic TN were included in this analysis. Although caution is warranted when approaching a patient with other possible reasons for facial pain, MVD may be considered when symptoms suggest classic TN and/or microvascular compression is strongly suggested on the MRI.^49,73 Fourth, given that all studies were retrospective, they are vulnerable to potential biases, which included interviews/surveys by nonindependent administrators and recall and response bias. Publication bias may have skewed the results toward more favorable outcomes. It is possible that some variables were underreported in the literature because of a lack of outcome prediction, which would bias the effect sizes observed.

CONCLUSION

This systematic review and meta-analysis described rates and predictors of pain freedom following MVD for TN. MVD provides an excellent long-term outcome for patients with TN, with approximately 76% remaining pain free with a mean follow-up of 1.7 yr. Positive predictors of favorable outcome included disease duration shorter than or equal to 5 yr, arterial compression, SCA involvement, and type 1 facial pain classification. The results will help preoperative patient counseling prior to MVD in this debilitating facial pain disorder.

Disclosures

DJ Englot is funded by the NIH/NINDS (1R00-NS097618). JD Rolston is supported by the NIH/NCATS 1KL2TR002539-01. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

Neurosurgery Speaks!  Audio abstracts available for this article at www.neurosurgeryonline.com.

COMMENTS

Microvascular decompression can be very effective for trigeminal neuralgia, but it is well known that some patients do better than others. In this article, the authors performed a detailed meta-analysis of 46 studies (7 prospective) to determine rates and predictors of pain freedom after microvascular decompression (MVD), defined as BNI score 1 (pain-free off medications). Overall, 76% of 3987 patients achieved pain freedom after a mean follow-up period of 1.7 yr. Significant predictors of success included disease duration under 5 yr, arterial compression, superior cerebellar artery involvement, and Burchiel Type 1 classification. The authors conclude that an understanding of these characteristics may have important implications for patient selection, preoperative counselling, and provider expectations.

The results of this study confirm the spectacular efficacy of MVD and support the widespread consensus that patients with arterial compression and Burchiel Type 1 symptoms are most likely to improve. The fact that shorter duration of symptoms prior to surgery is associated with better outcome is significant because it suggests that earlier surgery might lead to superior results. As with most meta-analyses, the study is somewhat limited by heterogeneity of outcome measures across the different studies and by evolution in technique, especially with respect to preoperative MRI sequences. The information, nonetheless, has great value for preoperative assessment and prognosis.

Jonathan P. Miller

Cleveland, Ohio

The authors of the paper analyzed the last 30 yr of published studies on microvascular decompression for trigeminal neuralgia and came to conclusions that the procedure works for a vast majority of patients. They confirmed the common notion of certain predictors of success (shorter duration of disease, presence of arterial compression, etc), whereas dominance of constant pain, and other features consistent with Burchiel type 2¹ presentation, was shown to be a predictor of less favorable outcome. Interestingly, the findings that link outcomes with both duration of disease and the atypical neuralgia features may be taken as an indirect support of our old hypothesis that the natural history of trigeminal neuralgia includes gradual transition from Type 1 to Type 2,² making it less responsive to decompression, and therefore, an earlier surgical intervention should be considered to achieve better long-term pain relief.

Unfortunately, the analysis also indicates that almost a quarter of TN patients undergoing MVD are not pain-free after 1.5 yr following their surgery, which is quite disappointing because the usual immediate post-MVD pain relief rate is usually quoted as better than 90%, or even 95%.

I doubt that, based on data presented here, the surgery will be offered less often to patients with prolonged history of trigeminal neuralgia or those with other “negative” predictors, but having this set of data will definitely help in gauging the expectations in most clinical situations.

Konstantin Slavin

Chicago, Illinois

Neurosurgery Speaks (Audio Abstracts)

Listen to audio translations of this paper's abstract into select languages by choosing from one of the selections below.