Self-Reported Trismus: prevalence, severity and impact on quality of life in oropharyngeal cancer survivorship: a cross-sectional survey report from a comprehensive cancer center

Abstract

Objective

The purpose of this study was to estimate prevalence/severity of self-reported trismus, determine association with quality of life (QOL), and examine clinical risk factors in a large population of patients treated for oropharyngeal cancer.

Materials and methods

A cross-sectional survivorship survey was conducted among patients who completed definitive treatment for oropharyngeal carcinoma, disease-free ≥ 1-year post-treatment (median survival, 7 years among 892 survivors). Associations between trismus and QOL were also analyzed using MDASI-HN, EQ-5D, and MDADI. Dietary and feeding tube status were also correlated to trismus status.

Results

Trismus was self-reported in 31%. Severity of trismus positively correlated (r = 0.29) with higher mean interference scores reflecting a moderate association with quality of life (p < 0.0001). There was a negative correlation for MDADI composite scores (r = − 0.33) indicating increased perceived dysphagia related to trismus severity (p < 0.0001). EQ-5D VAS scores were also negatively correlated with trismus severity (r = − 0.26, p < 0.0001). Larger T-stage (p ≤ 0.001), larger nodal stage (p = 0.03), tumor sub-site (p = 0.05), and concurrent chemoradiation (p = 0.01) associated with increased prevalence of trismus. Diet negatively correlated (r = − 0.27) with trismus severity (p = < 0.0001), and survivors with severe trismus were also more likely to be feeding tube-dependent.

Conclusion

Severity of trismus appears to negatively impact quality of life and associate with various adverse functional outcomes in long-term oropharyngeal cancer survivorship. Trismus remains associated with advanced disease stages, tumor sub-site (tonsil), and addition of chemotherapy. Further investigation is merited for the dose-effect relationship to the muscles of mastication.

Introduction

Trismus, defined as the inability to fully open the mouth, is a common oral morbidity and neuromuscular toxicity associated with tumor, radiation, and surgical injury in head and neck survivors. Current estimates of oral morbidities such as trismus and its impact on quality of life (QOL) are primarily derived from small study populations, often with heterogeneous groups of head and neck cancer for whom treatment scenarios and radiation volumes vary significantly [1]. There is limited focused research surrounding the associ-ation between trismus and quality of life, particularly using a large, homogeneous group of oropharyngeal cancer (OPC) survivors.

Difference in trismus assessment styles, research designs, and definitions of trismus has led to a wide range of published prevalence estimates ranging from 0 to 100% depending upon the author and evaluation criteria [2–5]. While it is generally accepted that 5 to 38% of pa-tients with head and neck cancer will develop trismus, there is more recent evidence to suggest a larger prevalence; 38–42% [6, 7]. This is thought to be further complicated by the large variation in “normal” inter-incisal opening reported in the literature. Agerberg et al. [7] reported that normal opening ranged from 42 to 77 mm for men and 39 to 75 for women. For example, a 20% decrease for a patient with an initial opening of 70 mm is much less impactful than in a similar patient with an initial opening of 45 mm. With such a wide variability in inter-incisal opening, possibly related to factors such as sex, body stature, and age, some authors have recommended a measurement method that is proportionate to the patient’ s body size [8, 9]. Zawawi et al. [9] suggested that finger widths are a convenient index for assessing inter-incisal opening, reporting that 100% of healthy study participants (with normal mouth opening) are able to place at least 3 fingers widths inter-incisally (index, middle, and ring) between the central incisors. Aside from the pragmatic and proportionate advantages of this method, the “finger-width” method of measuring oral opening also facilitates self-assessment by the patient outside of the clinical environment.

Trismus has implications on quality of life and functional status in various domains, but these associations need to be quantified. Aside from the obvious limitation in oral hygiene, chewing, eating, and decreased voice resonance, patients suffering from trismus have limited ability for dental restoration and maintenance that more broadly might adversely impact QOL and health. This puts patients at increased risk for further dental/oral challenges [10].

Prevention can reduce oral morbidity presentation and severity. While Kamstra et al. [11] showed some benefit to early intervention, no clinical recommendations could be made due to large variability in opening and stretching techniques. As with most prevention trials, it is best to identify the most at- risk group. While some authors fail to identify clinical predictors for the development of trismus [12], others report a series of potential risk factors, such as the female gender [13, 14], non-drinkers [5, 14], larger tumor size, and location, especially the oropharynx [10, 13, 15], concurrent chemotherapy [16, 17] to be associated with increased risk of trismus. Small study population size and heterogeneous study populations and short investigation periods often limit these conclusions; larger, homogeneous study populations have been recommended [18].

The purposes of this study were to:

1. Estimate the prevalence and severity of self-reported trismus in a large survivorship survey

2. Characterize associations between patient-reported trismus and QOL using MDASI-HN (MD Anderson Symptom Inventory-Head and Neck Module), ED-5D, and MDADI (MD Anderson Dysphagia Inventory).

3. Correlate self-reported trismus with dietary restrictions and feeding tube-dependence.

4. Examine potential clinical risk factors for trismus to determine high-risk subgroups of OPC survivors.

Materials and methods

Approval for this study was granted by the institutional review board of The University of Texas, M.D. Anderson Cancer Center (MDACC), PA11-0936. This was a cross-sectional survivorship survey that was conducted on oropharyngeal cancer patients who received treatment at MDACC between 01/01/2000 and 12/01/2016. Adults, ≥ 18 years of age, with no evidence of distant metastasis, recurrence, or second primary were included in this study. Subjects with ≤ 1 year of follow-up were excluded from the study to avoid any potential overlap with the acute side effects of treatment; also, it has been determined that the most significant decrease in oral opening occurs within the first 9 months post-radiation therapy, at a rate of 2.4% decrease per month. After 12 months, the percentage decrease in oral opening is about 0.2% per month [19].

Participants were contacted via multiple modes of communication: either email via Qualtrics or myMDAnderson (a secure, personalized patient website) or US postal service via first-class mail with a return envelope. Subjects were asked to complete the survey once with reminders to non-responders. The survey was administered using an adaptation of Dillman’s method, including the following: (1) a letter of invitation via the US postal service to eligible patients 2–3 weeks prior to survey administration; (2) the survey questionnaire to all eligible participants via an online server or US postal service; and (3) two reminders to non-responders via US postal service at 2–3 weeks and 4–5 weeks after the initial contact.

The survey included a single item reflecting self-reported mouth opening, specifically the number of fingers would fit inter-incisally. Patient who reported being able to insert 3 finger widths or greater were considered to have normal opening, while patients reporting < 3 fingers were considered to have reduced oral opening or trismus. For the purposes of this study, mild/moderate trismus was considered as the ability to insert 2 finger widths inter-incisally, and 1 finger width or less was considered severe trismus.

The electronic medical records of the consented patients were reviewed, and pertinent data was extracted. Risk factors were selected from the literature and tested for association with trismus incidence via chi-square analysis. Patient demographic factors such as sex, age, and smoking history were included. Smoking use was divided into “never,” defined as total use for less than a year; “former,” defined as having quit > 1 year ago; and “current,” defined as currently using tobacco or having stopped < 1 year ago. Disease-related factors such as tumor sub-site, T-stage, N-stage, and HPV positivity were evaluated. Treatment-related factors such as therapeutic combination and radiation type were analyzed: specifically, 3-D conformal, intensity-modulated radiation therapy (IMRT) and proton therapy (IMPT). Because the majority of subjects were treated primarily with radiation therapy, a subgroup was created to analyze the effect of chemotherapy on the development of trismus to determine the effect of induction and concurrent chemotherapy. The effect of neck dissection as part of the treatment was also calculated.

All patients who undergo radiation therapy at MDACC undergo a patient education session and are routinely referred to oral oncologists and speech pathologists for proactive rehabilitation planning. As part of a proactive swallowing exercise regimen, swallowing exercises targeting laryngeal, pharyngeal, and jaw range of motion are routinely instructed by speech pathologists to patients receiving oropharyngeal radiation. During the time period under investigation, within these exercises, patients were taught to stretch the mouth, opening, and holding for 5 seconds; they are further instructed to do 10 repetitions, four times a day. A subgroup of subjects had documentation of patient-reported compliance to these exercises. Full compliance was defined as exercises were as prescribed during and/or after radiotherapy; partial compliance was defined as exercises were performed less frequently than prescribed, or non-adherent indicating no chart documentation of exercise practice. A chi-square was preformed to determine the influence on trismus development.

All statistics were calculated by Stata Data Analysis Software (Version 14.0 College Station, TX) and are reported by p values, with p ≤ 0.05 considered statistically significant. Descriptive statistics were tabulated and plotted. Statistical comparisons included chi-square tests for categorical outcomes and t tests for continuous measures with Spearman’s rho to estimate size of correlations between patient-reported outcomes (PRO) and trismus severity.

Patient-reported outcome instruments

The PRO instruments used for this study were as follows: MDASI-HN, EQ-5D, and MDADI. These instruments were selected based on validity and reliability within the HNC population.

The survey included the MDASI-HN (MD Anderson Symptom Inventory-Head and Neck Module), a psychometrically validated 28-item questionnaire wherein there are 6 questions specifically targeted to evaluate how much the symptoms interfere with activities of daily life [20]. The 6 interference questions are rated from 0 “did not interfere” to 10 “interfered completely.” The mean interference score is felt to be a surrogate for health-related quality of life and overall patient function. The six questions specifically pertain to enjoyment, mood, relationship, walking, working, and general quality of life.

The EQ-5D is a well-documented tool for assessment of an individual’s health, wherein the subjects provide an overall impression of their health on a visual analog scale (VAS) from 0 to 100, where 100 represents their best health imaginable [21].

The MDADI is a validated, 20-item questionnaire that assesses swallowing-related QOL. The scores are normalized from 20 (low-functioning) to 100 (high-functioning). The MDADI is comprised of 2 summary scores: a global and composite score. The global score is a single question assessing the overall impact of swallowing on quality of life. The composite score is a summary of the remaining 19 questions, as a weighted average of the physical, emotional, and functional subscales [22].

Dietary assessment

The participants were asked to assess their diets and place them into one of 6 categories: (1) NPO (nothing by mouth); (2) strict liquid diet; (3) pureed diet; (4) solid diet with restrictions; (5) regular, with liquid assistance; (6) regular diet. This was then correlated to trismus severity and a chi-square calculated. Also, history of feeding tube utilization during or after cancer treatment was assessed and descriptive statistics calculated.

Results

A total of 1728 participants met the inclusion criteria above in order to be eligible for the survey; 989 subjects responded for an initial response rate of 57.8%. Eighty-two subjects were eliminated for ineligibility due to recurrent, second primary or distant metastasis after further data screening, and additional 15 subjects were eliminated due to missing trismus responses; 892 subjects were included in the study. A total of 31% reported trismus in some form; 252 subjects (28%) reported mild/moderate trismus defined as at least 2 finger widths inter-incisally and 26 subjects (3%) reported severe trismus, defined as 1 finger width inter-incisally or less. Median cancer free survival time was 7 years with minimum of 2 years and maximum of 16 years.

Trismus in relation to clinical and demographic factors

Demographic risk factors such as sex (p = 0.18), age (p = 0.44), smoking history at diagnosis, and use at survey (p = 0.53, 0.91, respectively) did not significantly associate with trismus. Risk factors associated with trismus were either disease-related, i.e., larger T-stage (p < 0.001), larger N-stage (p = 0.03), sub-site within the oropharynx (tonsil > base of tongue, p = 0.02), or treatment-related, such as radiation type (p ≤ 0.001) and addition of chemotherapy to radiation therapy (p = 0.01). Of the 431 subjects in which patient-reported compliance was documented, 327 subjects were found to be, at least, partially compliant with the provided exercises. Within this compliant group, 73% (n = 240) reported no trismus (3 fingers or greater) while in the non-compliant group only 56% (n = 58) reported no trismus. Adherence to pre-treatment exercises (p = 0.003) appeared to significantly associated with decreased trismus prevalence. Full details are listed in Table 1.

Table 1.

Risk factor analysis for trismus

	N	No of patients (n = 892)		p value
Characteristic	892	Trismus − (3 fingers) n = 614 (69%)	Trismus + (≤ 2 fingers) n = 278 (31%)	Significance (≤ 0.05)
Sex
Male	755	513 (68%)	242 (32%)	0.18
Female	137	101 (74%)	36 (26%)
Age
< 60	220	159 (72%)	61 (28%)	0.44
60–69	405	275 (68%)	130 (32%)
≥ 70	267	180 (67%)	87 (33%)
Tobacco use (at diagnosis)
Never	413	292 (71%)	121 (29%)	0.53
Former	420	283 (67%)	137 (33%)
Current	59	39 (66%)	20 (34%)
Tobacco use (at survey)
No	857	589 (69%)	268 (31%)	0.91
Yes	31	21 (68%)	10 (32%)
T-classification
T1	328	255 (77%)	73 (23%)	< 0.001*
T2	344	237 (69%)	107 (31%)
T3	131	77 (59%)	54 (41%)
T4	89	45 (51%)	44 (49%)
N-classification (AJCCv7)
N0	83	55 (66%)	28 (34%)	0.19
N1	127	91 (72%)	36 (28%)
N2a	107	82 (77%)	25 (23%)
N2b	398	276 (69%)	122 (31%)
N2c	148	92 (62%)	56 (38%)
N3	29	18 (62%)	11 (38%)
N-classification
N0, N1, N2a, b	715	504 (70%)	211 (30%)	0.03*
N2c, N3	177	110 (62%)	67 (38%)
Tumor site
Tonsil	413	268 (65%)	145 (35%)	0.05*
Base of tongue	437	317 (73%)	120 (28%)
Soft palate	8	8 (100%)	0 (0%)
Pharyngeal wall	6	3 (50%)	3 (50%)
Glossopharyngeal sulcus	10	7 (70%)	3 (30%)
NOS	18	11 (61%)	7 (39%)
HPV status (n = 380)
HPV−	58	38 (66%)	20 (34%)	0.15
HPV+	322	211 (66%)	111 (34%)
Therapeutic combination
Radiation alone	268	190 (71%)	78 (29%)	0.05*
Induction w/ rad alone	131	99 (76%)	32 (24%)
Concurrent alone	309	207 (67%)	102 (33%)
Induction and concurrent	155	94 (61%)	61 (39%)
Surgery and rad alone	6	6 (100%)	0
Surgery and concurrent	10	8 (80%)	2 (20%)
Surgery alone	8	7 (88%)	1 (12%)
Radiation followed by salvage surgery	1	0	1 (100%)
Radiation followed by salvage chemo	4	3 (75%)	1 (25%)
Radiation type
3-d conformal	49	22 (45%)	27 (55%)	<0.001*
IMRT	694	479 (69%)	214 (31%)
IMPT	22	17 (77%)	5 (22%)
PORT	16	14 (88%)	2 (13%)
Other	103	75 (73%)	28 (26%)
Chemotherapy
No	282	203 (72%)	79 (28%)	0.17
Yes	610	411 (67%)	199 (33%)
Concurrent chemo/rad
No	418	305 (73%)	113 (27%)	0.01*
Yes	474	309 (65%)	165 (35%)
Induction chemotherapy
No	606	421 (69%)	185 (30%)	0.55
Yes	286	193 (67%)	93 (33%)
Neck dissection
No	668	459 (69%)	209 (31%)	0.89
Yes	224	155 (69%)	69 (31%)
Adherence to pre/during treatment exercises
No/minimal	104	58 (56%)	46 (44%)	0.003*
Partial/full	372	240 (73%)	87 (27%)
Unknown	132	89 (67%)	43 (33%)

IMRT, intensity-modulated radiation therapy; IMPT, intensity-modulated proton therapy; PORT, post-operative radiation therapy

Trismus in relation to survival time

Subjects with more severe inter-incisal opening overall had higher mean survival time (10 years) when compared with those subjects with no trismus (7 years) and mild trismus (8 years), suggesting that trismus may progress over time. Complete details are presented in Fig. 1.

Fig. 1.

Trismus severity in relation to survival time. Severe trismus is associated with increased survival time. Each error bar is constructed using 1 standard error from the mean

Trismus and quality of life as measured by symptom interference or EQ-5D VAS

Mean MDASI interference symptom scores were statistically significantly higher among survivors with self-reported trismus (p < 0.0001) and positively correlated (r = 0.29) with severity of trismus reflecting a decreased quality of life in survivors with moderate or severe versus those with mild trismus or no trismus. Complete details are presented in Fig. 2. The EQ-5D VAS scores are presented in Fig. 3. The mean ± SD EQ-5D VAS score for patients with no trismus was 83.48 ± 14.91, while patients with mild/moderate trismus reported scores of 77.10 ± 17.67 and severe trismus reported 60.56 ± 25.35 (p > 0.0001, r = − 0.26), likewise indicating that increased severity of trismus correlated with lower overall quality of life. In regard to MDADI scores, all mean scores were statistically significantly lower (p < 0.0001) and negatively correlated with severity of trismus indicating decreased perceived swallowing ability as presented in Fig. 4.

Fig. 2.

Mean interference scores in relation to severity of trismus

Fig. 3.

Mean ED-5D scores in relation to trismus severity

Fig. 4.

MDADI scores in relation to trismus severity

Dietary assessment

Dietary information was self-reported in 857 subjects. Participants without trismus were more likely to report eating a regular diet with no restriction (47%), while only 31% of subjects with mild/moderate trismus reported regular diets. Zero subjects with severe trismus reported regular diets; the majority of participants with severe trismus reported eating a pureed/liquid diet or were NPO, 37.5%. Complete results are presented in Table 2.

Table 2.

Dietary restrictions in relation to trismus severity. There is a statistically significant difference (p < 0.0001) and negative correlation (r = − 0.27) indicating that severity of trismus is correlated more dietary restrictions. MIO: no trismus = 3 fingers, mild/moderate trismus = 2 fingers, severe trismus = 1 finger width or less

Diet (n = 857)	No trismus (n = 592)	Mild/moderate trismus (n = 241)	Severe trismus (n = 24)
Regular, no restrictions (n = 358)	283 (47%)	75 (31%)	0 (0%)
Regular, liquid assist (n = 326)	231 (39%)	89 (37%)	6 (25%)
Solid with restriction (n = 145)	70 (11%)	66 (27%)	9 (38%)
Pureed/non-chewable (n = 4)	3 (1%)	0 (0%)	1 (4%)
Liquid (no solids) (n = 12)	4 (1%)	5 (2%)	3 (13%)
NPO (nothing by mouth) (n = 12)	1 (1%)	6 (3%)	5 (21%)

Of the subjects with severe trismus, 88% had a history of feeding tube placement compared with 61% of those with mild/moderate trismus and 39% of subjects without trismus. Complete results are presented in Table 3.

Table 3.

Feeding tube placement in relation to trismus severity. There is a statistically significant difference (p < 0.00001) and a positive correlation (r = 0.92) indicating that severity of trismus is correlated with feeding tube placement. MIO: no trismus = 3 fingers, mild/moderate trismus = 2 fingers, severe trismus = 1 finger width or less

Feeding tube placement (n = 892)	No trismus (n = 614)	Mild/moderate trismus (n = 252)	Severe trismus (n = 26)
Yes	238 (39%)	154 (61%)	23 (88%)
No	376 (61%)	98 (39%)	3 (12%)

Discussion

Trismus is highly prevalent and impactful in head and neck cancer survivorship. Across a wide domain of outcomes, these survey data support the functional and quality of life implications of trismus in long-term oropharyngeal cancer survivorship. Significant, moderate correlations were detected between trismus and symptom interference (MDASI-HN), quality of life (EQ-5D), and perceived dysphagia (MDADI). Trismus, also, appears to be progressive over time. Subjects with severe trismus (1 finger width or less) had the largest mean survival time (10 years) when compared with those with no or mild/moderate trismus (7 and 8 years, respectively). Trismus may become more prominent as a late term oral morbidity in this population. This may be particularly impactful for those with HPV-related malignancies due to increased treatment response, decreased recurrence rates, and increased survivability [23–25].

Prevention of oral morbidities is critical in cancer care. In regard to trismus, the literature to support preventative exercises for risk reduction of trismus has similar issues with small sample sizes and heterogeneous study populations, and therefore, there is no current prevention recommendation [11]. Prior to radiation therapy, patients at some centers are commonly given exercises that aim lessen morbidities associated with cancer therapy. Survivors that were at least partially adherent to jaw stretching exercise prior and during treatment were less likely to develop trismus. While this is not conclusive evidence that mouth stretching prior to and during radiation therapy is an effective means of prevention, however, these observational data support further investigation on this topic.

The risk factors identified in this study as statistically significant, e.g., tumor size and sub-site, type of radiation therapy, and concurrent chemoradiation, were either treatment-related or disease-related. An unexpected finding was that the majority of subjects who underwent proton therapy (IMPT) and post-operative radiation therapy (PORT) developed trismus: 77% (n = 17) and 88% (n = 14), respectively. Although these groups are small, this is an interesting avenue of future research, particularly in regard to IMPT as it is often thought of as less damaging to tissues. When the two most common sub-sites were analyzed, subjects with tonsil malignancies were more likely to develop trismus than those with a base of tongue malignancies. Malignancies of the tonsil have more superior target volumes that often include the muscles of mastication, specifically the medial and lateral pterygoid. Goldstein et al. [26] in an early publication determined that radiation to the pterygoid region was sufficient to cause limited mouth opening. Direct damage to these muscles is most likely implicated in the increased development of trismus. Other demographic risk factors, i.e., age, sex, and tobacco habits, were not significantly associated with trismus.

The addition of chemotherapy in a concurrent setting had a statistically significant relationship with the development of trismus. The mechanism for this relationship is unknown even though this has been suggested by other authors [16]. It is possible that the addition of chemotherapy to radiation therapy in a concurrent setting can increase fibrosis in critical muscles of mastication-causing trismus. An alternative explanation would be that larger/more advanced tumors are more likely to undergo concurrent chemoradiation versus radiation therapy alone, and thus, trismus may be a result of the larger radiation volumes and not the addition of chemotherapy.

While the gold standard for inter-incisal measurement is generally in millimeters, the use of finger widths has been reported [9]. While these methods may provide a more objective method of determining trismus prevalence rates, there are still limitations. For example, subjects with multiple missing anterior teeth, or missing all teeth, were not controlled for; specifically, patient with missing anterior teeth may be able to insert more fingers between the edentulous ridges, and thus, the rate of trismus may be underestimated. Also, in spite of being given directions on how to measure using finger widths with picture and text instructions (Appendix Fig. 5), this method can be open to misinterpretation or subject errors. The broader health implications of trismus in long-term survivors may not be elucidated by this work given that PRO instruments were selected with head and neck cancer outcomes focus in mind at the time the survey was compiled.

Conclusion

Trismus impacts 31% of patients in long-term survivorship after curative treatment for oropharyngeal cancer. Trismus severity is associated with overall decreased quality of life using EQ-5D and MDASI symptom inventory as a surrogate QOL. There is an increase in the symptoms of dysphagia which is correlated with trismus severity and supported by patient-reported dietary limitations. Patients with larger tumor size, advanced N-stage, and tonsil malignancies are more likely to develop trismus, while other like sex, age, and social habits did not appear to influence trismus prevalence.