Frey syndrome is a common complication after parotidectomy. The time from surgery to disease onset may be quite long; therefore, a time-to-event analysis was performed for the occurrence of this syndrome post-parotidectomy. Three hundred and thirty-four patients who underwent a parotidectomy between January 2002 and November 2012 were identified (retrospective study). Of these patients, 102 developed Frey syndrome post-surgery and 232 did not. The time-to-onset analysis enabled us to estimate the risk ratio associated with different types of parotid gland tumours, various parotidectomy procedures, and repeat parotidectomy, which is useful for predicting preoperative and surgical risk. The risk of developing Frey syndrome was lower in patients with malignant tumours than in those with benign tumours (risk ratio 0.351, 95% confidence interval (CI) 0.155–0.594). Risk ratios for lumpectomy PA (pre-auricular area), superficial parotidectomy, and total parotidectomy with respect to lumpectomy T (tail) were 4.378 (95% CI 1.168–16.410), 8.040 (95% CI 3.286–19.670), and 8.174 (95% CI 3.076–21.723), respectively. Repeat parotidectomy also increased the risk of developing Frey syndrome (risk ratio 3.214, 95% CI 1.547–6.678). No effect of the use of a superficial muscular aponeurotic system (SMAS) flap on the risk of developing Frey syndrome was detected ( P = 0.888).
A parotidectomy is the main course of treatment for primary neoplasms of the parotid gland. The following three surgical techniques are described for the removal of such neoplasms, depending on the location, size, and histological characteristics of the tumour: lumpectomy (tail of parotid gland (T), or pre-auricular area (PA)), superficial parotidectomy, and total parotidectomy.
Frey syndrome, also known as auriculo-temporal syndrome, is one of the most common post-surgical complications associated with parotidectomies, and minimizing the incidence of this syndrome is an important objective for surgeons. The main symptom is sweating over the pre-auricular and/or temporal regions several minutes after chewing and ingesting food; the sweating can also be accompanied by redness, heat, or pain in the same regions.
Koch et al. suggested that the amount of glandular tissue removed during parotidectomy is one of the most important risk factors for the development of Frey syndrome. The amount of tissue removed is typically related to tumour malignancy. Superficial and total parotidectomies are typically used to treat benign tumours of the superficial and deep lobes, respectively, because they are associated with the lowest risk of tumour recurrence. The SMAS flap is frequently used to improve the patient’s appearance and prevent Frey syndrome after parotidectomy ; however, the success of this technique in preventing the syndrome is controversial.
In this study the occurrence of Frey syndrome after parotidectomy was examined via a time-to-event analysis. The survival analysis is the most typical application of a time-to-event analysis; however, in this case, the onset of Frey syndrome was the event of interest.
The objective of this study was to assess the effects of the histological classification of the tumour (i.e., malignant or benign), the type of parotidectomy, and the use of a SMAS flap on the occurrence of Frey syndrome by analysing the time to syndrome onset after surgery.
Materials and methods
For this retrospective study, appropriate patients were identified through information in the clinical databases of the study institution. This work was performed in compliance with the principles stated in the Declaration of Helsinki Ethical Principles for Medical Research Involving Human Subjects, adopted by the 18th World Medical Assembly, Helsinki, Finland, June 1964, and as amended most recently by the 64th World Medical Assembly, Fortaleza, Brazil, October 2013. As this was a retrospective study, ethical approval was not required. Three hundred and thirty-four patients who underwent surgery for parotid tumours between January 2002 and November 2012 were identified in the database investigation.
After informed consent was obtained, the patients completed a questionnaire about the presence or absence of symptoms associated with Frey syndrome (e.g., sweat, pain, heat, and facial flushing). For affirmative responses, the date of symptom onset was recorded. Some of the surgical patients had undergone a previous parotidectomy. The starting time for such patients was defined as the date of the second parotidectomy, which was considered the cause of possible occurrences of Frey syndrome.
Of the 334 patients in this study, 102 showed symptoms of Frey syndrome (30.5%) and 232 did not (censored data group). Censored data were associated with patients who died prior to the survey (22 deaths, 6.6%), individuals for whom correct contact information was missing (26 unreached patients, 7.8%), and individuals who were free of symptoms associated with Frey syndrome at the end of the study (184 patients withdrawn on 2 November 2012; 55.1%). In all cases of death, the cause of death was unrelated to the parotid gland pathology. Based on the medical records of the patients who died, were unreachable, or were withdrawn from the study, we could estimate the time period during which they were free of symptoms associated with Frey syndrome.
Table 1 shows the demographic details, clinical characteristics, and co-morbidities of the study population. In the SMAS elevation group, the skin flap was raised in the subcutaneous plane superficial to the SMAS. The SMAS and contiguous platysma muscle were elevated as a separate layer. If the tumour involved the SMAS layer, the affected portion was excised with the specimen and the defect was repaired at closure, taking advantage of the fact that the SMAS is an elastic tissue. After tumour excision, the SMAS flap was reapproximated and sutured to the tragal plane. In this study, the SMAS flap technique was applied when the SMAS flap integrity remained intact.
|Variable||Censored data ( n = 232)||Development of Frey syndrome ( n = 102)||P -value|
|Age at intervention, years||52.79 (1.155)||47.63 (1.572)||0.011|
|Male||117 (50.4%)||42 (41.2%)|
|Female||115 (49.6%)||60 (58.8%)|
|Spain||218 (94.0%)||100 (98.0%)|
|Rest of Europe||5 (2.1%)||0 (0%)|
|Asia||2 (0.9%)||2 (2.0%)|
|Africa||7 (3.0%)||0 (0%)|
|Active smoking status||0.037|
|Smoker||105 (45.3%)||33 (32.4%)|
|Non-smoker||127 (54.7%)||69 (67.6%)|
|Yes||37 (16.0%)||17 (16.7%)|
|No||194 (84.0%)||85 (83.3%)|
|Hypertension (HTN) a||0.120|
|Yes||53 (22.8%)||15 (14.7%)|
|No||179 (77.2%)||87 (85.3%)|
|Diabetes mellitus type 1 (IDDM) b||1.000|
|Yes||3 (1.3%)||1 (1.0%)|
|No||229 (98.7%)||101 (99.0%)|
|Diabetes mellitus type 2 (NIDDM) c||0.146|
|Yes||15 (6.5%)||2 (2.0%)|
|No||217 (93.5%)||100 (98.0%)|
|Yes||7 (3.0%)||0 (0%)|
|No||225 (97.0%)||102 (100.0%)|
|Yes||25 (10.8%)||9 (8.8%)|
|No||207 (89.2%)||93 (91.2%)|
|Yes||21 (9.1%)||4 (3.9%)|
|No||211 (90.9%)||98 (96.1%)|
|Yes||23 (9.9%)||4 (3.9%)|
|No||209 (90.1%)||98 (96.1%)|
|Yes||7 (3.0%)||10 (9.8%)|
|No||225 (97.0%)||92 (90.2%)|
|Type of parotid gland tumour||0.305|
|Malignant||26 (11.2%)||7 (6.9%)|
|Benign||206 (88.8%)||95 (93.1%)|
|Type of parotidectomy f||<0.001|
|Lumpectomy T||74 (31.9%)||7 (6.9%)|
|Lumpectomy PA||17 (7.3%)||4 (3.9%)|
|Superficial||107 (46.1%)||71 (69.6%)|
|Total||34 (14.7%)||20 (19.6%)|
|Yes||73 (31.5%)||39 (38.2%)|
|No||159 (68.5%)||63 (61.8%)|
Statistical analyses were performed according to the principles set forth by Lang and Secic. The χ 2 test (alternatively Fisher’s exact test) was used to compare proportions and the t -test to compare means between the patients with Frey syndrome and the censored data group. The symptom-free rate was estimated using the Kaplan–Meier method (i.e., the product-limit method). The Breslow generalized Wilcoxon test was used to compare the probabilities of remaining symptom-free. The Cox proportional hazards regression analysis was used to assess the associations between explanatory variables and the symptom-free rate; a forward stepwise method based on the Wald statistic was used for variable selection. IBM SPSS Statistics v. 19.0 software (IBM Corp., Armonk, NY, USA) was used for all of the statistical analyses.
The diagnostic and follow-up procedures did not change throughout the study period. The proportional hazards assumption in the Cox regression analysis was met. Censored data occurred uniformly over the follow-up period. We could expect that the distribution of consequences was identical for the unreachable patients and patients remaining in the study.
Figure 1 illustrates the development of Frey syndrome over time after parotidectomy. The mean symptom-free time was 1551.1 days (standard error (SE) 231 days).
Significant differences ( P < 0.05) between the Frey syndrome and censored data groups were observed for age at intervention, active smoking status, repeat parotidectomy, and type of parotidectomy ( Table 1 ).
Table 2 shows the distribution of adjuvant therapy used for cases with malignant parotid gland tumours; no significant difference was observed between the Frey syndrome and censored data groups (Fisher’s exact test for no treatment vs. treatment: P = 0.670).
|Variable||Censored data ( n = 26)||Development of Frey syndrome ( n = 7)||P -value|
|No treatment||14 (53.8%)||5 (71.4%)|
|Treatment (all types)||12 (46.2%)||2 (28.6%)|
|Type of treatment|
|Radiotherapy||8 (30.8%)||2 (28.6%)|
|Chemotherapy||1 (3.8%)||0 (0%)|
|Radio and chemotherapy||3 (11.5%)||0 (0%)|