Eleftherios Savvas, Steffen Hillmann, Daniel Weiss, Mario Koopmann, Claudia Rudack, Jürgen Alberty. Association Between Facial Nerve Monitoring With Postoperative Facial Paralysis in Parotidectomy. JAMA Otolaryngol Head Neck Surg. 2016;142(9):828–833. doi:10.1001/jamaoto.2016.1192
Electrophysiologic facial nerve monitoring is becoming an established intraoperative aid to assist the surgeon in facial nerve trunk or branch location and dissection. Limited studies have addressed the postoperative outcomes of parotid surgery with and without monitoring.
To examine the influence of intraoperative facial nerve monitoring on postoperative facial nerve function and procedure duration in parotid surgery.
Design, Setting, and Participants
An 8-year retrospective review of parotidectomies performed at the Department of Otorhinolaryngology–Head and Neck Surgery, University of Münster. The study analyzed 120 patients undergoing parotidectomy without monitoring from January 1, 1988, to December 31, 1991, and 147 patients undergoing parotidectomy with monitoring from January 1, 2003, to December 31, 2006. The patients were further subdivided in partial parotidectomy (PP) (n = 222) and total parotidectomy (TP) (n = 45) groups. An evaluation of operative time was performed to test the hypothesis of shorter duration of surgery with facial nerve monitoring. Final follow-up was completed on December 31, 2008, and data were analyzed from June 1 to December 31, 2013.
Main Outcomes and Measures
Comparison of the incidence of facial nerve dysfunction and operative time between the PP and TP subgroups with and without monitoring.
A total of 267 patients (127 men [47.6%] and 140 women [52.4%]; mean [SD] age, 51.3 [17.6] years; range, 3-90 years) were included in the analysis. A significant reduction in postoperative facial nerve dysfunction with the use of nerve monitoring could be seen in the PP group (46 of 99 without monitoring [46.5%] vs 18 of 123 with monitoring [14.6%]; P = .001). A similar finding was evident in the TP group when comparing moderate and severe nerve dysfunction (9 of 21 without monitoring [42.9%] vs 2 of 24 with monitoring [8.3%]; P = .01). The mean (SD) operative time in the PP subgroup without nerve monitoring was 115.3 (37.8) minutes; with nerve monitoring, 110.1 (33.6) minutes. The mean (SD) operative time in the TP subgroup without nerve monitoring was 134.5 (50.4) minutes; with nerve monitoring, 158.3 (56.3) minutes. There was no statistical difference between these groups.
Conclusions and Relevance
Facial nerve monitoring in primary parotid surgery for benign and malignant disease does not necessarily reduce the operative time, but the rate of transient postoperative facial nerve dysfunction or the grade of palsy is reduced.
The first report of a parotidectomy with preservation of the facial nerve was described in 1825 by Johann Ferdinand Heyfelder. Alfred Armand Louis Marie Velpeau successfully identified the facial nerve trunk in 1830. Intraoperative cranial nerve monitoring (including cranial nerve VII) has been used routinely during surgery for acoustic neuroma and on the middle fossa for many decades.1,2 The first reports of facial nerve monitoring in parotid surgery were published in the early 1990s.3- 6 Parotid surgery has evolved during this long time span. Modern surgical techniques aim for minimal surgical incision with an appropriate visualization of the surgical field, tumor removal with a margin of safety, and a low morbidity with an emphasis on facial nerve function. To accomplish these goals, electrophysiologic facial nerve monitoring is becoming an established intraoperative aid to assist the surgeon in facial nerve trunk or branch location and dissection. In a postal survey to all UK consultant otolaryngologists,7 82% of respondents performing parotid surgery reported monitoring facial nerve function during the procedure. In a similar US survey,8 60% of respondents who performed parotidectomy used facial nerve monitoring some or all of the time. In a nationwide survey of ear, nose, and throat hospitals in Germany in 2006,9 75% of respondents routinely used facial nerve monitoring in parotid surgery.
Publications addressing the postoperative outcome of facial nerve function after parotid surgery with and without facial nerve monitoring are limited and heterogeneous, often with small patient samples. Rates of postoperative facial nerve dysfunction depend not only on surgical expertise but also on tumor size, location, and differentiation. Scarred or inflamed tissue, as in chronic or recurrent disease and in revision surgery, is more susceptible to nerve damage during dissection of the parotid gland. These factors make reporting operative outcome difficult to classify and compare. Temporary facial nerve paralysis occurs in 15% to 66% of cases after primary parotid surgery, with slightly higher rates for total parotidectomy (TP) than for partial parotidectomy (PP). Permanent facial nerve paralysis is less common and occurs in 2.5% to 5.0% of cases.10 In revision surgery, the rates are higher owing to tissue fibrosis and are reported in as many as 100% of procedures in the immediate postoperative phase.11
In the present study, we undertook an 8-year review of parotidectomies performed in our institution from 1988, when intraoperative facial monitoring was not yet the standard of care. We compared changes in outcomes concerning postoperative facial nerve paralysis by overall rate and by the grade of dysfunction. Furthermore, we evaluated operative times to test the hypothesis that facial nerve monitoring results in a shorter duration of surgery. Although the setting up and testing of the monitoring equipment requires extra time, this time should be offset by the aid of the device if it helps the surgeon operate confidently at a quicker pace.
Question Does intraoperative facial nerve monitoring affect postoperative facial nerve function and procedure duration in parotid surgery?
Findings In this retrospective cohort study of patients undergoing partial and total parotidectomy, postoperative facial nerve dysfunction was reduced with the use of nerve monitoring in both groups. No significant difference in procedure duration was noted between groups.
Meaning Facial nerve monitoring in primary parotidectomy for benign and malignant disease does not necessarily reduce operative time but decreases the rate of transient postoperative facial nerve dysfunction or the grade of palsy.
The introduction and standardized use of facial nerve monitoring in parotid surgery was incorporated in the Department of Otorhinolaryngology–Head and Neck Surgery at the University of Münster during 1994 to 1995. In this study, we performed a retrospective review of data on all patients who underwent parotid surgery before the introduction of facial nerve monitoring from January 1, 1988, to December 31, 1991, and compared these data with those of all patients who underwent surgery with monitoring from January 1, 2003, to December 31, 2006. By 2003, the surgeons had gained enough experience with the monitoring device to be cognizant of possible setup errors and had used the device on a routine basis. The surgeons in both groups (including C.R. and J.A.) were long-time faculty members equally experienced in head and neck surgery. This study was approved by the institutional review board of the University of Münster. Owing to the retrospective nature of this study and the use of deidentified data, the institutional review board of the University of Münster waived the requirement of informed consent.
Data collection included tumor size, histologic diagnosis, surgical procedure, postoperative facial nerve function, and duration of operation. The surgical procedure was divided into a PP group that included superficial lateral operations and a TP group that included subtotal operations wherein part of the medial lobe was resected. Both groups were then subdivided into subgroups with and without monitoring.
The duration of the operation was defined as the time from incision to suture documented by the attending anesthesiologist. The setting up of the monitoring equipment was not recorded or factored into the operative time. However, such setup with the placement of electrodes and testing usually requires less than 5 minutes of preparation time. Analysis of facial nerve function was based on the daily notes of the resident or the surgeon for the preoperative, postoperative day 1, and final evaluations. We used a modified House-Brackmann score12 to classify the facial nerve function. To minimize the subjective bias of the observer encountered in the patient medical records, we chose to group House-Brackmann scores III and IV and scores V and VI together. We therefore had a scoring system of levels I to IV of dysfunction, with I indicating none; II, mild; III, moderate; and IV, severe.
Exclusion criteria consisted of a preoperative facial nerve weakness or paralysis, revision operations, combined operations extending to the neck (such as a neck dissection), and cases with facial nerve reconstruction. None of our patients required sacrifice of nerve branches (eg, owing to infiltration of malignant disease).
All procedures were performed in a standardized manner under general anesthesia. Magnification of the surgical field using a surgical microscope or loupes depended on the surgeon’s preference and was not further evaluated. In the procedures without facial nerve monitoring, a resident visually monitored facial nerve function and informed the surgeon of muscle movement or twitching. In the monitored group, a nerve monitor (CLEO; Inomed Medizintechnik GmbH) was used in addition. A 2-channel electromyography system was used with subdermal monopolar electrodes placed in the ipsilateral orbicularis oris and oculi muscles. A subdermal ground electrode was placed paranasally. Electromyographic changes in the corresponding mimic muscles were then monitored. Mechanical manipulation of the facial nerve evokes a high-frequency burst of motor unit potentials.13 These bursts can be heard and seen through the monitoring device, warning the surgeon. The mechanical stimulus and the muscle contraction are directly correlated, and once the stimulus ceases, the contraction ends. In addition, a testing probe for direct nerve stimulation is present with a stimulus range from 0.5 to 1.0 mA. All procedures were performed by attending surgeons (including C.R. and J.A.), senior-level residents, or fellows under supervision.
In all patients (in both periods), parotidectomy was performed using standard techniques. An anterograde preparation technique was typically performed in which the main trunk of the facial nerve was located and the peripheral branches followed for PP. The entire lateral part of the parotid gland was resected with the tumor. On occasion, when the tumor covered the main facial trunk, a retrograde or a combined dissection approach was necessary. In cases with tumor lying medial to the peripheral nerve branches, TP was necessary. In such cases, the retromandibular vessels were ligated and the deep lobe of the gland was resected.
Statistical analysis, conducted from June 1 to December 31, 2013, was performed using SPSS software (version 13.0; SPSS Inc). We used a 2-tailed t test to compare the duration of surgery. Data are expressed as mean (SD). Differences were considered significant at P < .05. Based on the 267 patients who underwent parotidectomy at our institution during the inclusion times, we had an 80% power to detect a difference of 14.3 minutes between monitored and unmonitored surgical procedures.
In the 2 study periods, a total of 267 patients underwent parotidectomies that met our inclusion criteria. Of these, 222 patients underwent PP and 45 underwent TP. Overall, 127 men (47.6%) and 140 women (52.4%) constituted the study group; 120 patients underwent parotidectomy without facial nerve monitoring, and 147 underwent parotidectomy with monitoring. The mean age was 51.3 (17.6) years (age range, 3-90 years); by sex, the men had a mean age of 50.1 (16.7) years; the women, 52.4 (18.4) years. Age showed a normal curve of distribution.
The mean tumor size was 2.47 (1.13) cm (range, 0.7-7.0 cm) with a normal curve of distribution. The duration of follow-up varied widely from 1 to 727 (median, 7) days. Two hundred thirty-nine patients (89.5%) underwent their final evaluation within 20 days after the operation; the few remaining patients showed a wide variability of follow-up time. Final follow-up was completed on December 31, 2008. The pathologic finding of the tumors was benign in 211 cases (79.0%), with benign disease most often pleomorphic adenoma (n = 132). Seventeen patients (6.4%) had malignant disease, and 39 patients (14.6%) had inflammatory or other disease.
In the PP group, 123 patients (55.4%) underwent intraoperative facial nerve monitoring and 99 patients (44.6%) did not. The mean ages in the subgroups with and without nerve monitoring were 53.3 (17.4) and 50.6 (17.2) years, respectively. Age in both groups showed a normal curve of distribution without any significant difference. Mean tumor size was 2.20 (0.94) cm in the subgroup with monitoring and 2.57 (1.27) cm in the subgroup without monitoring for a difference of 0.37 (95% CI, 0.08-0.67) cm. The pathologic findings showed a similar distribution of benign, malignant, and inflammatory disease in both PP subgroups.
In the TP group, 24 patients (53.3%) underwent intraoperative facial nerve monitoring and 21 patients (46.7%) did not. The mean age in the subgroup with monitoring was 47.3 (21.7) years; in the subgroup without monitoring, 50.7 (16.1) years. Age in both subgroups showed a normal curve of distribution without any significant difference. Mean tumor size was 3.05 (1.04) cm in the monitored group and 2.93 (1.07) cm in the unmonitored group for a difference of 0.12 (95% CI, −0.52 to 0.76) cm.
Facial nerve function was assessed in the PP group on postoperative day 1, with the following findings: no nerve dysfunction (grade I) was observed in 53 patients without monitoring (53.5%) vs 105 with monitoring (85.4%); mild dysfunction (grade II), in 27 without monitoring (27.3%) vs 16 with monitoring (13.0%); moderate dysfunction (grade III), in 16 without monitoring (16.2%) vs 2 with monitoring (1.6%); and severe dysfunction (grade IV), in 3 without monitoring (3.0%) vs 0 with monitoring. In the final evaluation (mean duration of follow-up, 20.4 [65.8] days for patients without monitoring and 23.7 [90.6] days for patients with monitoring), improvement with no dysfunction was found in 63 without monitoring (63.6%) vs 112 with monitoring (91.1%); mild dysfunction, in 21 without monitoring (21.2%) vs 9 with monitoring (7.3%); moderate dysfunction, in 16 without monitoring (16.2%) vs 2 with monitoring (1.6%); and severe dysfunction in no patient in either subgroup.
Statistical evaluation using multivariate logistic regression (independent variables, tumor size, and use of monitoring device) showed no correlation with tumor size (postoperative day 1, P = .26; final evaluation, P = .64). The use of the monitoring device reduced postoperative facial paralysis with an odds ratio of 4.81 (95% CI, 2.53-9.16) on postoperative day 1 and 5.93 (95% CI, 2.81-12.53) at the final evaluation.
Facial nerve function was assessed in the TP group on postoperative day 1, with the following findings: no nerve dysfunction (grade I) in 9 patients without monitoring (42.9%) vs 12 with monitoring (50.0%); mild dysfunction (grade II), in 3 without monitoring (14.3%) vs 10 with monitoring (41.7%); moderate dysfunction (grade III), in 7 without monitoring (33.3%) vs 2 with monitoring (8.3%); and severe dysfunction (grade IV), in 2 without monitoring (9.5%) vs 0 with monitoring. In the final evaluation (mean duration of follow-up, 14.0 [10.5] days for patients without monitoring and 51.7 [128.0] days for patients with monitoring), improvement with no dysfunction was seen in 10 patients without monitoring (47.6%) vs 17 with monitoring (70.8%); mild dysfunction, in 3 without monitoring (14.3%) vs 6 with monitoring (25.0%); moderate dysfunction, in 8 without monitoring (38.1%) vs 1 with monitoring (4.2%); and severe dysfunction, in no patient in either subgroup. The odds ratio of having a postoperative facial paralysis of grade III to IV in an operation without facial nerve monitoring was 8.25 (95% CI, 1.53-44.53) on postoperative day 1.
A total of 17 malignant tumors were found in the study. Patients with preoperative facial weakness were excluded in this study. Of these 17 patients, 4 (23.5%) had mild dysfunction on postoperative day 1 and 3 (17.6%), on final evaluation; 1 (5.9%) had moderate dysfunction on postoperative day 1 and final evaluation; the remaining 12 (70.6%) had no postoperative dysfunction.
The mean operative time for the PP group was 112.4 (35.5) minutes (range, 40-210 minutes). The mean operative time in the PP subgroup without nerve monitoring was 115.3 (37.8) minutes; in the subgroup with nerve monitoring, 110.1 (33.6) minutes. The difference in operative times between subgroups was 5.2 (95% CI, −14.7 to 4.3) minutes.
The mean operative time for the TP group was 147.2 (53.7) minutes (range, 70-315 minutes). The mean operative time in the TP subgroup without nerve monitoring was 134.5 (50.4) minutes; in the subgroup with nerve monitoring, 158.3 (56.3) minutes. The difference in operative times between subgroups was 23.8 (95% CI, −8.5 to 56.1) minutes.
This study aims to determine the effect of facial nerve monitoring on postoperative facial nerve function and the possible reduction in operative time. Although a randomized prospective study comparing similar cohorts with and without facial nerve monitoring would be the ideal construct, this seems hardly conceivable. As with all surgical comparisons, the groups would require similar tumor size, entity, and location; similar age and sex of the patients; and a similar level of expertise of the surgeons. In our retrospective analysis, these factors were similar in all 4 groups, with a significant difference only in tumor size of the PP subgroups with (2.2 cm) vs without (2.6 cm) monitoring (P = .01). The data represent a true difference in tumor size between the monitored and unmonitored subgroups as small as 0.08 cm to a maximum size difference of 0.67 cm. This size range is not clinically meaningful from a surgical standpoint; an increase in tumor size of as much as 1 cm should not change the surgical procedure or considerably displace the facial nerve. The tumor size did not show any correlation with postoperative facial nerve dysfunction. The factor of surgeon expertise is an important although subjective factor. The surgeons in both study periods were experienced senior-level colleagues; however, the surgeons in the period of 1988 to 1991 had been trained to operate without the aid of facial nerve monitoring. At present, fewer surgeons have experience without the monitoring aid and therefore establishing such a team would be difficult. These factors necessitate a retrospective analysis of patient data coupled to a review of the published literature on this topic.
A decrease in the rate of postoperative facial nerve weakness from postoperative day 1 was seen in our PP subgroups without and with monitoring at 46 patients (46.5%) compared with 18 patients (14.6%), respectively, and on final evaluation at 36 patients (36.4%) compared with 11 patients (8.9%), respectively. This difference was statistically significant (P = .001). Only a few studies have compared postoperative facial nerve function based on nerve monitoring, and the results are contradictory. In the reports of Deneuve et al14 and Grosheva et al,15 no significant improvement was found; however, Terrell et al16 showed a significant improvement in transient palsy. The TP subgroups without and with monitoring also had a decrease in postoperative facial nerve weakness at 12 patients (57.1%) compared with 12 patients (50.0%), respectively, and on final evaluation at 11 patients (52.3%) compared with 7 patients (29.2%), respectively. However, this difference was not statistically significant. A significant difference in the grade of palsy could be shown in the TP group only, with higher grades of palsy in the subgroup without monitoring (P = .01). This result and a faster recovery time in the monitored groups were seen by Makeieff et al11 and Liu et al17 in revision cases. Unfortunately, we cannot determine the rate of permanent facial nerve dysfunction in our patient pool. The recovery of facial nerve function can take 18 to 24 months.10,18 Of the 267 parotidectomies performed, 68 patients still had some form of facial nerve dysfunction within a 10-day postoperative period. Sixty of these patients were then lost to or unavailable for follow up; 2 patients returned with normal function; and 6 patients returned with continued paralysis. Owing to the large number of patients lost to follow-up relatively early, we cannot make any statement about permanent facial nerve dysfunction. Most of these patients had mild to moderate facial nerve weakness that perhaps was not debilitating enough for compliance with the follow-up examinations.
Several studies examining factors related to postoperative paralysis after parotid surgery have found age,10,16,19,20 operating time,16,20,21 and tumor size20,21 to be significant factors. Nouraei et al18 mentioned sialadenitis, multiple pathologic diagnoses, and parotid duct ligation, whereas Gaillard et al22 found tumor location, benign inflammatory lesions, and extent of parotidectomy to be significant negative factors in their patient pools. Such a factor analysis was not the primary goal of this study; however, the operating time, tumor size, and age did not show any correlation to facial nerve function in our population. Although such analyses may be of interest, one cannot change any of these factors except for surgery time. If surgery time is a significant factor, this situation asks whether facial nerve monitoring helps the surgeon expedite the operation. In our PP and TP groups, we found no statistical differences between the subgroups with and without monitoring. The small numbers of patients in the TP subgroups and the tumor entities may have affected the analysis. Several studies11,14- 17,21,23 have compared operative time in monitored and unmonitored groups; of these, 3 studies11,14,17 showed a significant reduction in operative time for TP; 2 studies,14,15 a significant reduction in time for PP; and 1 study,23 a significantly longer operative time in cases with chronic disease. Last, 2 studies16,21 showed no significant difference in the comparison. However, all studies showed a reduction in operative time (irrespective of statistical significance).
Two main benefits in the use of nerve monitoring have been noted. The first is locating the main trunk or branches of the facial nerve. Although landmarks help the surgeon, tumor infiltration of parotid tissue can alter the normal anatomy and make dissection difficult. The hypothesis of shorter duration of surgery with monitoring could not be validated in our study. Experienced head and neck surgeons trained without the use of the device had similar operating times as those trained with the routine use of nerve monitoring. An experienced surgeon will successfully find the nerve in a timely fashion, even when it is displaced. The second potential benefit of nerve monitoring is more apparent in our results: the immediate warning of inadvertent damage to the neural structure due to stretching, compression, or thermal damage. This damage may cause edema and impede the microcirculation, which blocks appropriate electric current potentials. This situation may explain the higher percentage of facial nerve dysfunction in the subgroups without monitoring. No authors, to our knowledge, have described any unwanted resection of nerve fibers without monitoring, by which we conclude that experienced surgeons will find the nerve with or without the aid. However, the striking consensus is that rates of more severe damage with delayed recovery are present in the unmonitored cohorts.11,17 This seems to be a result of the mentioned inadvertent damage when dissecting near a hidden nerve branch.
In the literature, transient facial nerve dysfunction varies from 6.5% to 100%, and permanent or long-term dysfunction, from 0.9% to 50.0%. However, comparing studies is difficult owing to differences in populations and standards, including differences such as type of surgery performed (enucleation, superficial, total), whether surgery is a primary or a revision procedure, tumor entity, and different reporting schemes (postoperative day 1, 1 week, 1 month, etc). Clinical assumptions that are validated in the studies include the fact that more extensive facial nerve dissection (as in PP vs TP and in revision surgery) leads to higher rates of transient facial nerve dysfunction but not necessarily permanent dysfunction.22- 25 Benign and malignant disease, when preoperative facial nerve function is normal and branches are not sacrificed for tumor control, have similar outcomes in facial nerve function. Benign inflammatory disease has a worse outcome owing to fibrosis of glandular tissue.18,22,26 A total of 17 malignant tumors (6.4%) were included in this study. The presence of malignant disease did not increase the rate of postoperative facial nerve dysfunction. A higher morbidity and mortality rate in parotid malignant neoplasms is usually associated with cT4 classifications at diagnosis and, as mentioned, with clinical involvement of the facial nerve.27,28 Such cases were not present in our cohorts.
Facial nerve monitoring does not replace a keen knowledge of facial nerve anatomy and appropriate surgical training. Surgeons must get acquainted with possible false-positive alarms, as when stimulating a vessel or fibrous tissue near a nerve branch, or when signal interference happens because 2 surgical instruments come into contact. Such artifacts are easily distinguished from true nerve stimulation with time.
Based on our results and a detailed review of the literature, electrophysiologic facial nerve monitoring in primary parotid surgery for benign and malignant disease, depending on the surgeon’s preference and experience, does not necessarily reduce the duration of surgery but does decrease the rate of transient postoperative facial nerve dysfunction or grade of palsy. This initial benefit seems to dwindle over time as nerve functions recover, leading to similar rates of permanent palsy in monitored and unmonitored subgroups.
Corresponding Author: Eleftherios Savvas, MD, Department of Otorhinolaryngology–Head and Neck Surgery, University of Münster, Kardinal von Galen Ring 10, 48149 Münster, Germany (email@example.com).
Accepted for Publication: April 25, 2016.
Published Online: June 16, 2016. doi:10.1001/jamaoto.2016.1192.
Author Contributions: Drs Savvas and Alberty had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Hillmann, Koopmann, Rudack, Alberty.
Acquisition, analysis, or interpretation of data: Savvas, Hillmann, Weiss, Alberty.
Drafting of the manuscript: Savvas, Hillmann.
Critical revision of the manuscript for important intellectual content: Savvas, Weiss, Koopmann, Rudack, Alberty.
Statistical analysis: Savvas.
Obtained funding: Alberty.
Administrative, technical, or material support: Rudack, Alberty.
Study supervision: Weiss, Koopmann, Alberty.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.