Functional outcomes and quality of life after a 6-month early intervention program for oral cancer survivors: a single-arm clinical trial

Early phase

The early phase intervention initiated at an average of 8.3 days following reconstructive surgery and lasted within one month post operation. In the early phase, the main goal was to help the participants deal with the problems suffered (i.e., pain, edema, shoulder dysfunction and strength loss, soft tissue tightness and functional limitation) as the result of the surgery. The intervention consisted were as follows: (1) Transcutaneous Electrical Stimulation was administered for 15 min. per treatment session and the intensity was adjusted to patient’s tolerance to alleviate shoulder pain or soreness. (2) Gentle soft tissue mobilization and scar massage on the donor and recipient sites was performed for 15 min. per treatment session and the intensity was adjusted to patient’s tolerance to prevent edema-induced stiffness. (3) To increase TMJ range of motion, the participants were guided to exercise their TMJ during the first visit and to use tongue depressors to assist in performing the TMJ exercise if needed in the following visits. (4) Active or active-assistive exercises for neck, shoulder and donor site were performed for 10 repetitions in each session. (5) The progressive shoulder resistance exercise consisted of closed-chain and open-chain exercises that were adjusted sequentially according to pain conditions (Fig. 1). (6) In cases involving fibular osteoseptocutaneous flap harvest, big toe flexion was avoided in the initial six weeks. (7) Functional training such as transfer, ambulation, and activities of daily living (ADL) was administered before discharge from the hospital. All of these were initiated as soon as the wound condition was stable (i.e., normal capillary refill decided by the surgeon clinically.) The participants were instructed to perform self-massage and other ROM exercises hourly during waking hours and was checked using exercise diaries kept by the participants. During the period of hospitalization (average duration = 3 weeks), participants received the intervention every weekday for an average of 40 min. per treatment session. The treatment program was administered once a week after discharge from hospital.

Middle phase

The middle phase lasted from 1–3 months post operation. During this phase, the goal was to focus on the impairment suffered (i.e., scar development, soft tissue tightness, shoulder joint dysfunction and strength loss, functional limitation) from the surgery or radiation therapy (RT). The participants attended the rehabilitation center for one-on-one therapy once per week. The interventions included: (1) Scar massage was employed 5 to 10 min prior to other treatment programs. (2) Soft tissue massage and joint ROM exercise as described in the early phase was implemented as well with intensity up to patient’s tolerance. Once the participants felt free of shoulder pain, the training was transferred to the open-chain exercise. (3) Progressive resistance exercise (PRE) training for the open-chain exercises was administered with free weights or thera-bands 1 set of 10 repetitions in one session (Fig. 2). In this phase, some of the participants started to receive RT, which can interfere with TMJ and oral function. (4) To maintain TMJ flexibility, conventional therapy for TMJ was maintained in these subjects. On top of that, oral function training was implemented as well. (5) Tongue and lip mobility and coordination training, and instructions on food intake were included. (6) In cases where a fibular osteoseptocutaneous flap was performed, the big toe and ankle joint were stretched in this phase to increase joint ROM and soft tissue flexibility. The training programs were evaluated and adjusted as needed during every visit. Each treatment session lasted for an average of 50 min. Participants were also instructed to perform individual home-programs on an hourly basis when awake at home. Compliance to home-programs was checked having the patients keep diaries regarding their home endeavors.

Late phase

The late phase started more than three months post-operation. The goal in this phase was to recover the residual functions as much as possible. The interventions included were: (1) Pain and scar management with intensity up to patient’s tolerance and (2) shoulder PRE programs were implemented as in the previous phase, i.e., 10 repetitions each session. (3) If microstomia was caused by the intra-oral scar contracture after RT, a microstomia splint was prescribed and intra-oral scar massage was applied. Treatments to improve oral, shoulder joint and physical functioning were performed and progressed according to individual performance. The participants attended the rehabilitation center once a week. Each treatment session lasted for 1 h during this phase. In addition to a weekly visit to the rehabilitation center, patients were asked to perform their individual home programs with hourly exercise of the TMJ and shoulder, and were also encouraged to perform physical exercise (i.e., scar massage, mouth opening exercise, shoulder-neck ROM and PRE, et al.) at least three to five times a week.

All the intervention program is based on a variety of clinical trials compiled in the review article by Guru, Manoor & Supe (2012).

All the treatments were provided by three certified therapists with an average of more than 10 years of clinical experience.

Baseline data included demographic information and outcome variables taken during the 1st visit after operation. Data at one, three and six months after the operation were also included. Furthermore, return-to-work status of all subjects employed at baseline were followed at 12-month post operation through telephone interview to confirm if they are back to work.

Outcome measurements

Demographic data including gender, age, marital status, education level, disease staging, oncology and reconstructive treatment were collected after informed consent was given. Outcome measurements included shoulder outcomes, oral health, health-related QoL and physical functions. Shoulder outcomes were evaluated according to shoulder joint ROM and muscle strength (MMT), the visual analog scale (VAS) for the assessment of shoulder pain, and the Disabilities of the Arm, Shoulder and Hand (DASH) Outcome Measure for upper extremity function. Oral health was evaluated using maximum mouth opening (MMO) and diet status, which was divided into nasogastric (NG) tube feeding, liquid diet, soft diet or normal diet. Health-related QoL was evaluated using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) and Quality of Life Questionnaire Core Head and Neck 35 (QLQ-HN35), and using physical function evaluations. Six-minute walking distance (6MWD) and the timed up and go (TUG) tests were employed to evaluate physical function.

MMT and ROM

The strength of the rotator cuff, pectoralis muscle, deltoid, trapezius, serratus anterior, rhomboid and latissmus dorsi were evaluated according to the numerical grading system of Medical Research Council (MRC) scale. A score of 4 is the acknowledged cut-off score that demonstrates the muscle’s ability to hold testing position against gravity and moderate resistance (Mark, 2017). Shoulder joint ROM was evaluated in flexion, abduction, external rotation, and internal rotation with a two-arm goniometer under standard procedures. The evaluation was terminated when any contraindications in the prone position were indicated such as breathing difficulty. Interincisor distance (IID) was collected in millimeters by a Willis gauge, in the case of healthy teeth. Maximal mouth opening (MMO) was measured by measuring the difference between the distances between the nose tip to mandible during mouth opening and mouth closing.

DASH outcome measure

The DASH outcome measure is a self-report questionnaire and a clinical tool to measure physical function and symptoms of the upper extremity, which studies have found useful for patients after neck dissection (Carr, Bowyer & Cox, 2009; Chan et al., 2015). The disability/symptom section contains 30 items scored from 1 to 5 indicating “no difficulty” to “unable” to perform the task. At least 27 items must be completed, scores are consequently transformed into a total scale of 1–100. Higher scores indicate greater disability. A minimum clinically important difference (MCID) was used to assess the smallest perceived important change (Cranney et al., 2001). A change score exceeding 15 points is the most accurate change score for discriminating between improved and unimproved state (Beaton et al., 2001).

EORTC QLQ-C30 and QLQ-HN35

The EORTC QLQ-C30 is a questionnaire developed to evaluate the general quality of life of cancer patients. The QLQ-H&N35 is one of the disease-specific module supplements for head and neck cancer. The QLQ-H&N35, in conjunction with the QLQ-C30, is considered a reliable and valid assessment of the quality of life among patients with head and neck cancer in various different countries (Bjordal et al., 2000; Chie et al., 2010; Scherman, Simonton & Adams, 2000). The EORTC QLQ-C30 contains 30 questions and is divided into a global health scale, five functional (physical, role, cognitive, emotional, and social) scales, and nine symptom (fatigue, pain, nausea/vomiting, dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties) scales. The EORTC QLQ-H&N35 contains 35 questions and is divided into six symptom scales (pain, swallowing, senses (taste/smell), speech, social eating, and social contact), and seven single items (impaired sexuality, teeth problems, mouth opening, dry mouth, sticky saliva, coughing, and feeling ill). Every scale is transformed into a score ranging from 0 to 100. A higher score on the functional scale or global health scale represents a higher level of functioning or quality of life. In contrast, a higher score on the symptom scale or single item scale reflects a worse symptom or problem.

Physical functions

The 6-minute walking distance (6MWD) and the timed up and go (TUG) tests were employed to evaluate physical capacity and lower extremity function. 6MWD is a tool used to evaluate functional exercise capacity of cancer patients and has been suggested to be as valid and reliable in this context as it is in evaluating cardiopulmonary patients (Schmidt et al., 2013). The test was performed in 6 min, and the patients were asked to walk to the end and back on a 20 m walkway as fast as they can, without overexertion. The total distance was recorded in millimeters. The TUG test is a commonly used tool that is evidently based to predict functional mobility and stability in less healthy, lower-functioning adults (Schoene et al., 2013). It measures the time in seconds to stand up from a sitting position on a chair, walk 3 m, turn, walk back, and to sit back down on the chair.

All the evaluations were conducted by a different group of two certified therapists blinded to the study, with an average 12.5 years of clinical experience.

Statistical analysis

SPSS 20.0 was used in the data analysis with a maximum significance level set at 0.05. Independent t-tests were used for the analysis of continuous variables. For the categorical data, chi-squared tests or Fisher’s exact tests were used to analyze differences in the independent variables of gender, marital status, educational level, breadwinner, vocation, TNM stage, neck dissection, radiation therapy, diet and donor site, while independent t-tests were applied to test for differences in the dependent variables of age, 6MWD and TUG. The Generalized Estimating Equations (GEE) procedure was conducted to analyze repeated measures outcome variables over time. We used the GEE model which assumed unstructured working correlation matrix. GEE has the benefit of a robust estimator, and it can overcome the limitation of missing data and adjust for correlations between observations as well. Separate models were run for each outcome variable. Bilateral shoulders were evaluated separately when bilateral neck dissections had been performed. The data were excluded if the participants had shoulder problems before the operation. Pearson correlation coefficient was used to measure the degree of association between outcome variables. Kaplan–Meier estimates were used for the survival analysis to predict the rate of return-to-work. The log rank test was conducted to compare differences in the rate of return-to-work between the early stage and advanced stage groups.

Study population

A total of 65 participants (60 male and five female, mean age 51.6 years) who underwent 80 neck dissections (15 subjects received bilateral neck dissections), were analyzed in this study after reconstructive microsurgery. Thirteen eligible subjects declined to enroll. The drop-out rate was 34%. The reason was decondition and fatigue caused by RT that decreased their willingness to participate in the intervention program. The CONSORT flow diagram is shown in Fig. 3. The demographic data are summarized in Table 1. A supplementary table (Table S1) is provided to present the demographic comparison between the drop-out and the analytic groups. To explore differences in the outcomes between the early stage and the advanced stage group, we compared them with respect to all variables of interest (using group model). Apart from the disease-related variables, there were no significant differences in the demographic data of the early staged group and the advanced group.

Shoulder outcomes

All participants had neck dissections, and none had a prior history of shoulder pain. With respect to shoulder ROM of the affected shoulder, only abduction showed a significant difference between six months and baseline or compared with the unaffected side during six months of the study (p < .001) (Tables 2 and 3). MMT of the middle trapezius, lower trapezius and rhomboid muscles were not performed at the baseline because patients were unable to comfortably maintain a prone posture. The muscle strength of the infraspinatus, subscapularis, teres muscles, pectoralis muscles and latissmus dorsi demonstrated normal strength at the baseline and exhibited no significant differences in the subsequent tests. Apart from the upper trapezius and levator scapular which gained stability of normal strength (4.9 ± 0.3) one month post operation (Table 2), the anterior and middle deltoid, trapezius, serratus anterior, rhomboid all showed significant differences in strength between the 6-month assessment compared to baseline, one month and three months, individually (p < .001) (Table 2). At one-month evaluation, only middle and lower trapezius scored less than 4 on the MMT scale. Other muscles attained sufficient strength to overcome some resistance and gravity. Comparison of the ratio of the shoulder outcomes of the neck dissection side and the sound side indicates that the shoulder ROM and muscle strength reached almost equivalent level to the sound side except in the middle and lower trapezius at six months of the study (Table 3).

The average shoulder pain score on the VAS was 1.81 ± 2.28 at the baseline, and 1.09 ± 1.89 at six months. VAS scores showed no significant difference between six month and baseline. However, the pain scale measured in the self-reported quality-of-life questionnaire EORTC QLQ-C30 showed a significantly higher score at the baseline (44.9 ± 26.0) (p < .001) compared with subsequent evaluations (Table 4). The mean DASH score at the baseline was 34.4 ± 24.0 and significantly dropped to 17.4 ± 16.2 at 1 month (p < .001), which was considered a minimum clinically important difference (MCID).

Oral health outcomes

At the baseline, a nasogastric (NG) tube was placed for all participants. One month after the operation, 18% persisted with NG tube feeding, 14% returned to normal diet, 28% and 40% started to get nutrition with a liquid and a soft diet, respectively. At six months, all NG tubes had been removed, and 53% returned to normal diet. 16% and 31% participants were able to orally intake a liquid and a soft diet.

IID and MMO had high positive correlation (r = .883). The mean IID was 22.5 ± 9.2 millimeters and significantly progressed to 31.8 ± 11.8 millimeters (p < .001) in the overall model. The mean MMO was 22.6 ± 2.1 millimeters (baseline) and 31.8 ± 10.1 millimeters (6 months), which was also a statistically significant difference (p < .001). In the group model, IID and MMO for both early stage and advanced stage groups significantly increased at 6 months compared to baseline, one and three months, individually as well. In the early stage group, IID increased from 21.8 ± 7.9 (baseline) to 34.0 ± 13.7 (six months) (p < .001), and it increased from 23.1 ± 10.1 (baseline) to 29.1 ± 8.6 (six months) (p = .010) in the advanced stage group.

Health related quality of life and physical functions

Table 4 illustrates the results of the EORTC QLQ-C30 scales. After six months of intensive physical therapy, health-related quality of life was significantly different (p < .05) from the baseline on the global health scale and all functional scales, except cognitive function. The symptom scales of fatigue, nausea/vomiting, pain, dyspnea and insomnia showed significant differences between the baseline and six-month scores or between three-month and 6-month scores (p < .05) for the overall and group models. However, no significant difference was found on the scales of constipation over time throughout the six months. While the scales of loss of appetite, diarrhea, and financial problems showed significantly high scores (p < .05) at baseline compared with 6-month scores in the early stage group, it maintained at similar level over time throughout the 6 months in the advanced stage group.

Table 5 illustrates the results from the EORTC QLQ-H&N35 scales. All other symptom scales showed a significant difference (p < .05) between the six-month test scores and the baseline scores in overall model and group model except on the scale on senses (tastes, smell) which showed no significant difference over six months in the early stage group. Oral pain, swallowing, and speech scores on the three-month test were significantly different (p < .05) compared with those on the six-month test in the advanced group. The single item scales of opening mouth, dry mouth, sticky saliva and coughing showed significant differences between the baseline and six-month scores, and between 3-month and 6-month scores (p < .05) except for the score on teeth and sexuality, which maintained at the same level over the 6 months in the early stage group and the advanced group.

At three-months and six-months, 6MWD was 339.0 ± 54.4m and 381.2 ± 69.9m, respectively (p < .001). TUG was 10.6 ± 2.4 s at 3 months and 8.4 ± 1.5 s at 6-months (p < .001). 6MWD on the 3-month and 6-month tests, respectively, was 355.5 ± 55.9m and 396.5 ± 73.5m (p = .029) in the early stage group, and 321.6 ± 48.2m and 363.1 ± 64.0m in the advanced stage group (p = .003). No significant difference was found between groups at 3 or 6 months on either measure.

Our analysis found a predicted return-to-work rate of 34.7% at three months, 63.0% at the six months and 80% at one year after surgery, respectively. The log rank test showed the time of return-to-work was significantly different between the early stage group and the advanced stage (p = .032). In the early stage group, it predicted that about 80% would return to work one year after the operation, but around 50% would return to work in the advanced stage group (Fig. 4). Further analysis showed 71.4% return-to-work rate in the retired or unemployed, 86.7% in the self-employed, 80% in the professional, 33.3% in the administration, 81.8% in the service and 87.5% in the semi-skilled or skilled according to vocational type.