Speech and Swallowing Impairment After Treatment for Oral and Oropharyngeal Cancer

Objectives  To assess the prevalence of speech and swallowing impairment after radical surgery for oral and oropharyngeal cancer from the patient's viewpoint and to examine the association of these functional alterations with selected clinical characteristics regarding patients, tumors, and oncologic treatment.

Design  Cross-sectional, multicenter study using a self-administered questionnaire.

Setting  Forty-three hospitals in Germany, Switzerland, and Austria.

Patients  A total of 3894 questionnaires about rehabilitation problems after treatment for oral and oropharyngeal squamous cell carcinoma were sent to patients. Of these, 1652 were filled out and returned, and 1334 (80.8%) met the inclusion criteria.

Main Outcome Measures  Morbidity associated with treatment of oral and oropharyngeal cancer.

Results  Speech problems were reported by 851 patients (63.8%), and swallowing problems were reported by 1006 patients (75.4%). The variables that presented a significant association with speech and swallowing impairment were sex, tumor location, pTNM stages, stage of tumor, treatment modality, and reconstruction type.

Conclusions  This survey, based on patient perception, suggests that those who undergo radiotherapy associated with the surgical removal of a tumor, have late-stage tumors (III-IV), or have tumors located in the floor of the mouth should be informed of the greater risk of persistent severe speech and swallowing problems.

The follow-up of patients operated on for oral and oropharyngeal cancer includes the evaluation of recurrent, persistent, and new primary cancers and the provision of psychological and social support to patients and family members. In addition, patients should be monitored for the development of nonneoplastic morbidity related to treatment.^1-3

Few studies^4-6 are based on functional outcomes as secondary end points in oral cancer treatment. Speech and swallowing are frequently rated as important by patients after primary surgery for oral and oropharyngeal cancer.^7-9 Although swallowing and speech are included in several scales of health-related quality of life in oral and oropharyngeal cancer,^10-12 few studies^7,8 have explored the predictive factors that influence these domains.

Detecting speech and swallowing impairment after oral and oropharyngeal cancer treatment would allow us to provide additional information to patients before treatment and to involve them in therapeutic decision making. Moreover, assessing posttreatment impairment could constitute an educational objective in the formation of the dental teams and their integration in multidisciplinary oncologic teams. This article aims to assess the prevalence of speech and swallowing impairment after radical surgery for oral and oropharyngeal cancer from the patient's viewpoint and to examine the association of these functional alterations to selected clinical characteristics regarding patients, tumors, and oncologic treatment.

A cross-sectional, multicenter study was performed under the auspices of the Deutsch-Österreichisch-Schweizerischer Arbeitskreis für Tumoren im Kiefer und Gesichtsbereich (DÖSAK), a German, Austrian, and Swiss study group for cancer in the maxillofacial region. A total of 3894 standard DÖSAK questionnaires about rehabilitation problems after treatment for oral and oropharyngeal squamous cell carcinoma were sent to patients from 43 hospitals in Germany, Switzerland, and Austria. All of these patients had undergone primary radical surgical treatment for oral and/or oropharyngeal cancer (lip cancer not included) at least 6 months earlier. The protocol for surgical treatment was based on DÖSAK criteria.¹³ These criteria are clinically intended safety margins of at least 1 cm for radical resection in 3 dimensions and removal of the ipsilateral and/or contralateral lymph nodes. Early-stage tumors were only treated surgically, whereas late-stage tumors often received a combined treatment of adjuvant radiotherapy and/or chemotherapy. Selection criteria for this study also included patients with nonrecurrence and no metastasis at the time of evaluation. The treating physician provided clinical data, such as age, sex, tumor location, pTNM status, oncologic treatment, type of reconstruction of the missing tissue (local tissue vs flap or bone graft), and time of survival free of tumor.

The standardized DÖSAK rehabilitation questionnaire^6,7 includes 147 items, divided into 9 relevant sections regarding morbidity associated with the treatment of oral and oropharyngeal cancer. The questionnaire covers areas such as personality and social background (10 items), medical history before the diagnosis of oral and oropharyngeal cancer (10 items), medical history from admission to surgery (11 items), medical admission after surgery (9 items), side effects from surgery and radiotherapy (21 items), side effects associated with daily living (24 items), feelings toward oral and oropharyngeal cancer (40 items), living conditions and circumstances (9 items), and lifestyle (13 items). It is a patient-based, self-administered questionnaire using a Likert-type scale. The scale of response consists of 5 categories ranging from very good or nonexisting impairment (rated as 0) to very bad or severe impairment (rated as 4). Sample characteristics are summarized in Table 1. The DÖSAK rehabilitation questionnaire^6,7 was used to retrospectively analyze patient perceptions regarding 3 stages of treatment (just before surgery, immediately after surgery, and at the time of completing the questionnaire). To do so, the following questions were repeated for each of the stages: (1) “To what degree did you experience impairment in speaking?” and (2) “To what degree did you experience impairment in swallowing?”

Statistical analysis was performed using a commercially available software package (SPSS release 12.0 for Windows; SPSS Inc, Chicago, Illinois). The descriptive statistical analysis was based on the mean and standard deviation for continuous variables, whereas the frequency and percentage were used for categorical variables. Two categories were established for speech and swallowing impairment: nonexistent or light (scored as 0 or 1) and moderate to severe (scored as 2-4). We applied the McNemar-Bowker test of symmetry to retrospectively compare speech and swallowing impairment in different stages of treatment. The association of the demographic, clinical, and treatment variables with speech and swallowing impairment were estimated using odds ratios and 95% confidence intervals by means of simple logistic regression methods. Afterward, multiple logistic regression methods were applied to obtain predictive models of speech and swallowing impairment based on covariates.

Of the questionnaires initially remitted, 1652 were filled out and returned. Inclusion criteria were met by 1334 patients (80.8%). Speech problems were reported by 851 patients (63.8%), whereas swallowing problems were reported by 1006 patients (75.4%). Statistically significant differences were found between frequency of speech and swallowing problems (P < .001). A total of 402 patients (30.1%) had a moderate to severe score for speech impairment (2-4 level on a Likert-type scale), and 748 patients (56%) had a moderate to severe score for swallowing impairment (2-4 level on a Likert-type scale).

The minimum time for obtaining the questionnaire was 6 months after treatment, and the maximum time was 162 months (mean [SD], 43.6 [33.6] months). Our statistical analysis revealed no relationship between speech impairment and time since treatment. However, we found a relationship between swallowing and time since treatment. Patients with more time elapsed since treatment present with more swallowing impairment.

When we retrospectively compared patients' perception of speech and swallowing impairment immediately before treatment, immediately after treatment, and at the time of completing the questionnaire, we found that impairment for speech and swallowing before treatment was better compared with after treatment (P < .001) and compared with at the time of completing the questionnaire (P < .001). We also found that speech and swallowing impairment was better at time of completing the questionnaire compared with immediately after treatment (P < .001). The distribution of speech and swallowing scores according to variables is summarized in Table 2 and Table 3.

Using a simple logistic regression model, all variables except age present a significant association with speech and swallowing impairment, although the significance levels vary. Men had worse functional outcome scores than women (speech, P = .01; swallowing, P = .001), tumors located in the floor of the mouth and the oropharynx obtained worse scores than other tumor locations (speech, P = .01; swallowing, P = .001), and smaller early-stage tumors resulted in better speech and swallowing scores (P = .001 for both).

Regarding treatment modalities, patients who received radiotherapy had the worst functional outcome (speech, P < .001; swallowing, P < .001), and with respect to reconstruction modalities, the less demanding reconstructive procedure (ie, primary closure or local tissue group) had the best functional outcome (speech, P = .04; swallowing, P < .001).

After applying multiple logistic regression methods, we obtained a predictive model of speech impairment with the following group of covariates: treatment (P = .001), location (P = .02), and stage (P = .02). The sensitivity and specificity of this model were 61.0% and 58.8%, respectively. A predictive model was also obtained for swallowing impairment, which included the following group of covariates: treatment (P < .001), stage (P < .001), and location (P = .004), with a sensitivity of 75.8% and a specificity of 51.4%.

This cross-sectional study presents the usual limitations of this type of analysis, for example, the lack of time sequence, a possible survival bias, and the influence of nonresponses. Nevertheless, the following steps were taken to reduce these biases: the clinical characteristics of nonresponders were verified to be similar to those of responders, deceased patients were not considered, and the likelihood of response was maximized by having questionnaires remitted by the medical departments where the patients were treated.

Analyzing oral function after radical surgery is complex because a variety of interacting variables are involved.¹⁴ The present study has focused on the most common sequelae of oral and oropharyngeal cancer treatment, which are speech intelligibility and swallowing problems. In the present study, swallowing impairment was significantly more prevalent than speech problems. This is in line with the work by Nicoletti et al,⁵ who reported a more complete recovery to speech than swallowing and chewing after oral and oropharyngeal cancer treatment. They speculated that this could be attributable to the higher cortical control of speech with respect to swallowing and chewing, which are largely reflex functions. Another explanation could be the tendency of patients to subjectively overestimate the quality of their own speech. The latter explanation, however, is not accepted by some authors.¹⁵ In any case, future study in this area would benefit from a more objective patient-based outcomes tool, such as the Dysphagia-Specific Quality of Life Questionnaire by McHorney et al,^11,12 to more accurately evaluate treatment variations and effectiveness.

To improve survival of oral and oropharyngeal cancer, several therapeutic strategies have been developed (surgery, radiotherapy, and chemotherapy). As a result, many patients with oral and oropharyngeal cancer have access to a range of treatment options that may differ not only in terms of survival but also in terms of acute morbidity, severity, and nature of long-term symptoms or impairments. Deleterious effects of radiotherapy have been described elsewhere.¹⁶ In our multivariate analysis, radiotherapy was the most important predictive factor related to both speech and swallowing impairment, perhaps because sensory changes and tissue fibrosis as a consequence of radiotherapy are factors that will affect neuromuscular coordination and the mobility of structures such as the remaining tongue, hyoid bone, and larynx.⁷ Nevertheless, Zuydam et al⁷ found that adjuvant radiotherapy was the most predictive factor for swallowing impairment only. Some studies^17,18 have even reported better functional outcome with primary radiotherapy treatment than with surgical resection, especially with respect to base of tongue and soft palate cancer. In any case, however, most authors^19-22 conclude that radiotherapy has a strongly negative effect on functional outcome; thus, indications for postoperative radiotherapy may need reconsideration.

Once oral and oropharyngeal cancer has been firmly established and before treatment, patients and relatives should be informed about immediate and late effects of each modality of treatment, explaining not only aesthetic sequelae but also functional deterioration. This should be remembered even though patients' concern for survival in the pretreatment phase often supports the use of aggressive treatment strategies.⁶

In agreement with other authors,⁵ when predicting impairment we have found that the key anatomical site in terms of functional outcome is the floor of the mouth followed by the posterior oral cavity. Loss of integrity in the genioglossus and geniohyoid muscles (important in providing tongue mobility) or the geniohyoid and mylohyoid muscles (fundamental for the normal elevation of the larynx) severely impairs speech quality and swallowing.

We found that patients with advanced tumor size or nodular affectation had significantly worse functional outcome. This may be the result of greater surgical removal, which worsens speech and swallowing impairment. Demanding reconstructive procedures were found to have worse functional outcome in a simple logistical regression analysis of our results. Nevertheless, in our multivariate analysis, reconstruction type is not a predictive factor for functional outcome. This is in contrast to the study by Zuydam et al,⁷ which found that the use of primary closure surgery was predictive of good speech and swallowing. Regarding reconstruction type, no consensus can yet be found in the literature.^23-26 According to Borggreven et al,²⁷ to obtain the best possible level of functional outcome, reconstruction should aim to maintain as much mobility as possible of structures in the oral cavity and oropharynx.

In conclusion, this survey, based on patient perception, suggests that those who undergo radiotherapy associated with the surgical removal of a tumor, have late-stage tumors (III-IV), or have tumors located in the floor of the mouth should be informed of the greater risk of persistent severe speech and swallowing problems. These findings may help the oncologic team (oncologists, surgeons, radiologists, dentists, psychologists, and speech therapists) to provide further details of treatment options and functional sequelae.

Correspondence: Nils-Claudius Gellrich, DDS, MD, PhD, Department of Oral and Maxillofacial Surgery, Hannover Medical School, D-30625 Hannover, Germany (Gellrich.Nils-Claudius@mh-hannover.de).

Submitted for Publication: April 11, 2007; final revision received January 19, 2008; accepted April 8, 2008.

Author Contributions: Drs Suarez-Cunqueiro, Schoen, Kokemueller, and Gellrich 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: Suarez-Cunqueiro, Schramm, Otero-Cepeda, Kokemueller, Metzger, Diz-Dios, and Gellrich. Acquisition of data: Schramm, Schoen, and Bormann. Analysis and interpretation of data: Suarez-Cunqueiro, Schramm, Seoane-Lestón, Otero-Cepeda, and Kokemueller. Drafting of the manuscript: Suarez-Cunqueiro, Seoane-Lestón, Bormann, Metzger, and Gellrich. Critical revision of the manuscript for important intellectual content: Suarez-Cunqueiro, Schramm, Schoen, Otero-Cepeda, Kokemueller, Diz-Dios, and Gellrich. Statistical analysis: Schramm, Otero-Cepeda, and Metzger. Administrative, technical, and material support: Schoen and Bormann. Study supervision: Suarez-Cunqueiro, Schramm, Seoane-Lestón, Kokemueller, Diz-Dios, and Gellrich.

Financial Disclosure: None reported.

Additional Contributions: We thank DÖSAK for its support, the participating patients, and the oral and maxillofacial surgery departments that contributed their patients, time, and efforts.