Long-term Outcomes and quality of life following parotidectomy for benign disease: A cohort study

Background: Parotidectomy may be burdened by numerous complications that may worsen subjects' quality of life (QoL). So far, the literature still lacks of long-term data (>10 years) answering to the question what impacted the patients the most on QOL after parotidectomy compared to well-published short-term data. Methods: A prospective long-term follow-up study was carried out. Participants were divided into three groups concerning the follow-up: short-term (ST; 6 postoperative weeks), long-term (LT; 13 years postoperative) and short-and long-term (SLT) on same patient collective. QOL was assessed by the Parotidectomy Outcome Inventory (POI-8). Demographic and clinical data were collected from all patients. Operative reports were used to classify all parotidectomies as great auricular nerve (GAN) “preserving” or GAN “sacri�cing” surgical preparations. Results: 74 LT, 57 ST and 33 SLT patients were enrolled in this study. Hypoesthesia posed the major short-and long-term problem whereas facial palsy posed the minor problem. Pain (p < 0.01) and hypoesthesia (p < 0.001) signi�cantly improved from six weeks to 13 years after parotidectomy as well as the overall POI-8 score (p = 0.04). The disease-specic impairment rate decreased from short (» 70%) to long-term (» 30%) follow-up. Sacri�ce of the auricular nerve was associated with hypoesthesia in the ST-cohort (p = 0.028). Conclusion: To our knowledge, this study represents the longest follow-up of patients undergoing parotidectomy. Hypoesthesia signi�cantly improved but still remains on long-follow-up without impacting QOL. As part of the preoperative informed consent, prolonged or permanent hypoesthesia should be explicitly emphasized.


Introduction
Tumours of the parotid gland constitute approximately 3% of the head and neck tumours, and about 80% of them exist as benign pathologies.Benign tumours like pleomorphic salivary adenoma, cystadenolymphoma or basal cell adenoma, in ammatory and autoimmune diseases are indications for surgery [1].The aim of parotid surgery in benign diseases is the partial or total removal of the gland while preserving the facial nerve function and preventing tumour recurrence [2].The overall complication rate is about 21.6% [3,4].Post-operative numbness, pain, dermal problems, mouth dryness and scar-related problems are the important complications that affect quality of life (QOL) [5].Health-related quality of life (HRQoL) measurements are gaining increased importance in clinical medicine.The Parotidectomy Outcome Inventory (POI-8) is the rst reliable and valid instrument for measuring health-related QoL in patients after parotidectomy for benign disease [6].Several studies reported no changes in QoL in a short observation period after parotidectomy -usually 6 months to 1 year [2,5,[7][8][9][10].However, little is known about the long-term disease-speci c QOL after parotidectomy.
The aim of this retro-and prospective study was to determine the long-term surgery-affected QOL using the disease-speci c POI-8 questionnaire and to compare those to well-published short-term data.In order to get more objectivity, data research was performed in different (LT-versus ST-cohort) and same patients collectives (SLT-cohort).

Material And Methods
The Ethics Committees of University of Heidelberg (Project Trial No: S-300/2007 and S-443/2018) granted approval for the study.This long-term follow-up study based on the validation data of the German version of the Parotidectomy Outcome Inventory-8 (POI-8) questionnaire including 199 patients undergoing benign parotidectomy [6] between 2003 and 2006 in our Department for Ear, Nose and Throat, Head and Neck Surgery, University of Heidelberg, Germany of which 57 (28,6%) patients returned the POI-8 six weeks after parotid surgery.Surgical procedures included partial and total parotidectomies.All histopathological results were retrieved, and patients with malignant diseases were excluded.The prospective recruitment took place in 2018.In total, 199 patients who were included in the original retrospective study of Baumann et al. [6] were reinvited to participate in the study of which 74 (37.2%) responded in 2018, this means 12-16 years after surgery.108 / 199 (54.3%) patients did not respond due to change of location or disinterest, 16 patients (8%) had already died and only one patient (0.5%) actively declined to participate in the study.For simpli cation, we assigned the study patients to different groups (Figure 1): patients who participated in our prospective study in 2018 tted into the long-term cohort (LT-cohort, 74/199) because they replied 12-16 years after parotidectomy.33 of them were already included in the original study of Baumann et al. in which the patients returned the POI-8 six weeks after parotid surgery [6].Consequently, these responders who returned the completed questionnaires six weeks and 12-16 years after surgery were assigned to the SLT-cohort (short-and long-term cohort, 33/199).Both groups were compared to the original short-term data (6 weeks postoperatively) of Baumann et al. [6], which were titled here as ST-cohort (short-term cohort, 57/199).Taken together, the pro-and retrospective POI-8 data from a total of 98 (74 LT, 57 ST, 33 SLT) patients were available.Hence, the LT-and ST-cohort are not disjoint but overlapping.The set of overlapping patients has the size of the SLT-cohort.All patients were informed about the study aims and protocol, and participants were enrolled after giving informed written consent.The Parotidectomy Outcome Inventory-8 (POI-8) is a validated and reliable questionnaire for patient-reported QOL in parotid surgery [6].It consists of eight Likert-type scaled questions from 0 to 5 (no -marginal -slight -moderate -severe -the worst problems) with a total score of 40.Low values for the POI-8 scales indicate high functionality and QoL.In addition to the POI-8, patients responded to global questions concerning the parotidectomy impairment (no -low grademoderate -severe -very severe impairment), their satisfaction with the postoperative results and if they would recommend this operation to good friends and family members.Additionally, questions were answered regarding doctor rounds, rehabilitative measures and days of incapacitation for work in the last 6 months.Group comparisons were made between the Short (ST)-and long-term (LT) cohorts (intergroup) as well as within the SLT-cohort (intragroup), short versus long follow-up measurements.
Lastly, operative reports and clinic notes were used to classify all parotidectomies as "nerve preserving" or "nerve sacri cing" surgical preparations concerning the branches of the great auricular nerve (GAN).

Statistical analysis
The data were analysed using R, Version 3.6.1.Metric variables are presented as means ± standard deviation, while factorial variables are presented by their proportional distribution.Additionally, t-statistics for the comparison of mean values between variables were used to determine differences between the cohorts (intergroup comparisons).Furthermore, the distribution of particular variables were tested for relatedness/independence with other variables, using -tests for intragroup comparisons.We decided on that particular procedure, since a) most of the variables that we compared are categorical and nominal distributed, b) the sample size was not too large to impact sensitivity of the test's results in a way that would lead to false positive outcomes and c) it is the statistic that is most commonly used for such comparisons, which may lead to more accessible and understandable results [11].A p-value less than 0.05 was considered statistically signi cant.The graphical displays that are shown, were programmed using the ggplot2 library of R.

Results
Patient cohorts 74 patients (41 men, 33 women) with an average age of 66.4 ± 12.2 years at time of survey prospectively completed the POI-8 questionnaire.Average time span of follow-up was 13.3 ± 1.1 (range: 12-16 years).The contributions of the benign histopathologic diagnoses were predominantly cystadenolymphoma (40.5%) and pleomorphic adenoma (28.4%), followed by chronically relapsing sialadenitis (6.8%) and parotid cyst (6.8%).93.2% of patients underwent a lateral, 6.8% of patients a total parotidectomy.Details of the different patient's cohorts and disease characteristics are shown in Table 1.
Table 1 Surgical and clinico-pathological characteristics of the three different study cohorts.Six weeks after parotidectomy, 77,2% of the ST-cohort characterized hypoesthesia as the most disturbing problem, followed by xerostomia (47.4%) and the appearance of the scar (45.6%).Facial palsy was posed as minor problem (87.7%).In the LT-cohort, hypoesthesia (54.1%), followed by fear of revision surgery (44.6%) and appearance of the scar (39.2%) posed the major problems.95.9% of the patients described facial palsy not as problematic.
42.1% of ST-cohort characterized pain on the site of surgery signi cantly more problematic with the mean POI-8 score of 0.93 than LT-cohort with the mean POI-8 score of 0.34 (p = 0.002).77% of LT-patients a rmed no pain or painful sensations 13 years after parotidectomy.Hypoesthesia improved signi cantly over the years but still posed a problem: 77.2% of ST-patients were disturbed by hypoesthesia six weeks after surgery with the mean POI-8 score of 1.86 and 54.1% of the LT-cohort still named sensation loss as a problem with the POI-8 mean score of 0.88 (p < 0.001).
In the ST-cohort, 64.9% of patients were -predominantly slightly -affected in disease-speci c QOL after parotid surgery.68.4% patients lled out to be "very satis ed" with operation results.In the LT-cohort, 27% of patients still described impairment after parotidectomy, thereof 4.1% as "severe".There was a longterm satisfaction rate of 81.8%.Negligible doctor visits (6.8%)/days of incapacitation (2.7%) and no rehabilitative measures were documented in the LT-cohort Intragroup differences in POI-8 outcomes in SLT-cohort Six weeks after parotidectomy, 84.8% of the SLT-cohort characterized hypoesthesia as the most disturbing problem, followed by the appearance of the scar (54.5%) and pain (51,5%).Facial palsy posed the minor problem (87.9%).At 13 years follow-up, hypoesthesia still bothered 60.6% of them, but only 3% severely, as well as Frey's syndrome (48,5%) and fear of revision (42,4%).No one suffered from facial palsy in the LT-cohort.In the meantime, from 6 weeks to 13 years after surgery, hypoesthesia (p < 0.001) and pain (p = 0.004) had signi cantly improved from patient's perspective.Dissatisfaction with the scar (p = 0.13), appearance due to substance loss (p = 0.17), Frey's syndrome (p = 0.36), xerostomia (p = 0.09), fear of revision surgery (p = 0.75) and facial palsy (p = 0.08) did not signi cantly ameliorate after surgery.The total score of POI-8 signi cantly decreased from the mean POI-8 score of 8.24 six weeks after surgery to the mean POI-8 score of 5.15 13 years after surgery, indicating higher overall satisfaction (p = 0.04; Figure 2).
Six weeks after parotidectomy, 72.7% of patients felt -mostly slightly -impaired by high satisfaction rate of 69.7%.In the long-term interval, 33,3% of patients felt impaired after parotid surgery, the satisfaction rate amounted 78.8%.No days of incapacitation, no rehabilitative measures and nearly no medical doctor visits (6.1%) were necessary in the last 6 months.
All group comparisons are graphically illustrated in Figure 3.

Preservation of the great auricular nerve (GAN)
Based on all available surgical reports (n = 98), 45.9% of the GAN could be technically preserved, 33.7% had been sacri ced and in 20.4% no information was found in the operation report.
All patients of the LT-cohort (n = 40; 23 men, 17 women), who documented a sensation loss in the POI-8 questionnaire, were contacted again and were asked about the location of the numbness.The area overlying the parotid gland and at the angle of the mandibule were slightly more affected than the area of auricle and along the anterior border of the sternocleidomastoid muscle (55% versus 45%).In general, the patients did not feel limited in QOL: the majority of the patients described the hypoesthesia as "marginal" (55%) or "slight" (35%), but 10% of them felt "moderately"/"severely" or "very severely" affected.Concerning this selected collective here (n = 40), there were no correlations between the nerve's sacri ce, the second POI-8 item, the operation procedure (lateral vs. total) and the affected area detectable (p > 0.05).Sacri ce of the GAN was only associated with the second item of the POI-8 (hypoesthesia) in the ST-cohort (p = 0.028), not in the SLT-or LT-cohort, using chi-square test.

Discussion
This study provides rst e cient data on short-(6 weeks) versus long-term (13 years) HRQoL after parotidectomy using the validated disease-speci c questionnaire POI-8.Key goal in parotid surgical approaches for treatment of benign lesions is not only to remove the entire tumour and to minimize morbidity but also to maintain patients' quality of life.However, only few studies address patients' perspective on complications and the related QoL [5,[7][8][9][12][13][14], but predominantly in a short follow-up after parotidectomy -usually 6 months or 1-2 years -and by using different symptom-speci c QoL assessments.Beutner et al. [8] reported no changes in QoL in 34 patients one year after super cial parotidectomy for benign diseases compared with preoperative answers using the EORTC QLQ-C30 and EORTC QLQ-H&N 35.Nitzan et al. utilized a modi ed version of the University of Washington Quality of Life Questionnaire and could detect subjective sequelae like altered sensation, change in appearance, Frey's syndrome and pain during at least 1-year follow-up period which did not signi cantly affect QoL [9].New data of Bulut et al. described for the rst time a signi cant increase in QOL in the long term (100 months postoperatively) compared to short term (2 weeks postoperatively) by focusing on sensory dysfunction resulting from great auricular nerve (GAN) sacri ce versus preservation in parotid surgery [14].Since most studies have only a short follow-up, it raises the question what impacted the patients the most on QOL more than 10 years after parotidectomy.To our knowledge, this is the longest follow-up reported after benign parotid surgery.From patient's perspective, hypoesthesia, followed by fear of revision surgery posed the major problems in long-term follow-up whereas facial palsy posed the minor problem.In the literature, numbness or uncomfortable sensations of the skin are well known as dominant, postoperative short-term sequelae after parotidectomy [5,7,9,12,15].Indeed, hypoesthesia is of greatest concern to the patients 6 months (90%) and still 2 years after surgery (78%) [7].Porter and Wood [16] observed that the majority of sensory improvement occurred in the rst 6 months.Ryan and Fee described that at a mean point of 2 years, symptoms had either completely ablated or stabilized [17].The great auricular nerve (GAN) originates from the cervical plexus at the levels of the second and third cervical nerves.It supplies sensation to the skin overlying the lower aspect of the pinna and angle of the mandible and is divided into anterior and posterior branches.The anterior branch is usually sacri ced whereas a posterior branch can technically be preserved [5].That's why, the area overlying the parotid gland and at the angle of the mandibule were slightly more affected here than the area of auricle and along the anterior border of the sternocleidomastoid muscle.In this context, Bulut et al. reported that GAN preservation did signi cantly improve sensation in short-but not in long-term, nor did it increase QOL in long term when compared to GAN sacri ce [14].Based on the observation period of 12-16 years after parotidectomy, we could show that hypoesthesia signi cantly improved over the years (p<0.001)but still remains without limiting patients in QoL, a clinically relevant nding here.Indeed, 77.2% of ST-patients were disturbed by hypoesthesia six weeks after surgery with an overall impairment rate of 64.9% while 54.1% of the LT-cohort named sensation loss still as a problem with an impairment rate of 27%.In the LTcohort, hypoesthesia was predominantly characterized as a "marginal" to "moderate" problem (51.3%) problem; only 2.8% of patients considered it as "severe" or "the worst problem".Consequently, patients should be preoperatively informed about the possible prolonged or permanent hypoesthesia.According to data of Bulut et al. [14], we only found a positive association between intraoperative preservation of the GAN and improved hypoesthesia in the ST-cohort (p = 0.028), not in the LT-cohort.Finally, we can conclude that patients seem to adapt to the postoperative functional impairment over time and focus less on the reduced ability to feel temperature and pain on the facial skin over the parotid gland and auricle.42.1% of the ST-cohort characterized pain on the site of surgery signi cantly more problematic than the LT-cohort (p = 0.002).Only 23% of LT-study patients a rmed pain and painful sensations 12-16 years after surgery, 13.5% of them described the pain as "marginal", 9.5% of them as "slight" and "moderate", no one as "severe" or as "the worst problem".Wolber et al. [7] and Nitzan et al. [9] underlined our results by reporting an incidence of postoperative pain of 30-40% during a short follow-up.In accord with Kaya et al. [10], we can consequently argue that pain is an important early complication following parotidectomy which improves in the long-term.Strikingly, the ST-and LT-cohort did not signi cantly differ in scar, facial palsy, substance loss, Frey's syndrome, xerostomia and anxiety of revision surgery.
Due to the fact that the total POI-8 score signi cantly improved from the mean POI-8 score of 7.47 six weeks after surgery to the mean POI-8 score of 5.15 13 years after surgery (p = 0.04), we can conclude that parotidectomy did not seem to be detrimental to QOL in the long run.Examinations on the same patient collective (SLT-cohort) underlined our results herein presented: hypoesthesia (p < 0.001) and pain (p = 0.004) had signi cantly improved from patient's perspective with a signi cant POI-8 reduction from 8.24 to 5.15 13 years after surgery, indicating higher disease-speci c QoL. Generally, global quality of life is a multidimensional construct with contributions from economic, social, interpersonal, physical and psychological aspects [8].Thus, subjective perception of post-parotidectomy complications might be in uenced by these mentioned factors unrelated to surgery [2] as disadvantage of self-reported measures.Besides, no objective testing method exists for accurate evaluation of QOL in the head and neck region [14].Another limitation of our study is the different group size in the ST-, LT-and SLT-cohort.Consequently, further prospective studies with a larger sample size are needed.In summary, we see the strength of our study in the rst reported, longest follow-up period of 12-16 years after parotidectomy, numerous group comparisons in different (LT, ST) and same patient cohorts (SLT_short vs. SLT_long), usage of a validated, commonly used patient-reported outcome measure POI-8 and in the retro-and prospective study design.

Conclusion
To our knowledge, this is the longest follow-up study assessing QOL in patients undergoing parotidectomy, which showed an overall improvement of QOL after surgery with a signi cant reduction of POI-8 total score 12-16 years after parotidectomy.From patient's perspective, hypoesthesia and the appearance of the scar posed the major long-term problems (SLT/LT-cohort) whereas facial palsy posed the minor problem.Even though, hypoesthesia signi cantly improved over the years, it still remains without impacting QOL and should be mentioned as part of informed consent.

Figure 1 Overview
Figure 1

Figure 2 Group
Figure 2

Figure 3 Group
Figure 3