- © 2002 by American Society of Clinical Oncology
Prophylaxis of Radiation-Associated Mucositis in Conventionally Treated Patients With Head and Neck Cancer: A Double-Blind, Phase III, Randomized, Controlled Trial Evaluating the Clinical Efficacy of an Antimicrobial Lozenge Using a Validated Mucositis Scoring System
- S. El-Sayed,
- A. Nabid,
- W. Shelley,
- J. Hay,
- J. Balogh,
- M. Gelinas,
- R. MacKenzie,
- N. Read,
- E. Berthelet,
- H. Lau,
- J. Epstein,
- P. Delvecchio,
- P. K. Ganguly,
- F. Wong,
- P. Burns,
- D. Tu and
- J. Pater
- From the Ottawa Regional Cancer Centre, Ottawa, Ontario; CUSE, Fleurimont, Quebec; Kingston Regional Cancer Centre, Kingston, Ontario; British Columbia Cancer Agency, British Columbia; Toronto-Sunnybrook Regional Cancer Centre, Toronto, Ontario; CHUM, Notre Dame, Quebec; Sudbury, Northeastern Ontario Regional Cancer Centre, Ontario; St. Jones, Newfoundland; CancerCare Manitoba, Manitoba; and National Cancer Institute of Canada, Clinical Trials Group, Kingston, Ontario, Canada.
- Address reprint requests to Samy El-Sayed, MD, University of Ottawa and Ottawa Regional Cancer Center, 503 Smyth Road, Ottawa, ON K1H 1C4, Canada; email: samy.el-sayed{at}orcc.on.ca
Abstract
PURPOSE: Mucositis occurs in almost all patients treated with radiotherapy for head and neck cancer. The aim of this multicenter, double-blind, prospective, randomized trial was to evaluate the clinical efficacy of an economically viable antimicrobial lozenge (bacitracin, clotrimazole, and gentamicin [BcoG]) in the alleviation of radiation-induced mucositis in patients with head and neck cancer.
PATIENTS AND METHODS: One hundred thirty-seven eligible patients were randomized to treatment with either antimicrobial lozenge (69 patients) or placebo lozenge (68 patients). The primary end point of the study was the time to development of severe mucositis from the start of radiotherapy. Secondary end points included severity and duration of mucositis, pain measurement, radiation therapy interruption, and quality of life. Mucositis was scored using a validated mucositis scoring system.
RESULTS: Toxicity profiles were similar between the two arms of the study. The median time to development of severe mucositis from the start of radiotherapy was 3.61 weeks on BCoG and 3.96 weeks on placebo (P = .61). There were no statistically significant differences between the arms in the extent of severe mucositis as measured by physician, in oral toxicities as recorded by patients, or in radiotherapy delays.
CONCLUSION: This study was conducted on the basis of a pilot study that demonstrated the BCoG lozenge to be tolerable and microbiologically efficacious. A validated mucositis scoring system was used. However, in this group of patients treated with conventional radiotherapy, the lozenge did not impact significantly on the severity of mucositis. Whether such a lozenge would be beneficial in treatment situations where rate of severe mucositis is higher (ie, in patients treated with unconventional fractionation or with concomitant chemotherapy) is unknown.
AN URGENT NEED exists for improved treatment of the oral complications of cancer therapy.1 There is currently no generally accepted and effective means to prevent or reduce treatment-induced mucositis.
Radiation mucositis is initiated because of the direct cytotoxic effect of radiation.2 Radiotherapy is directed toward cancer cells but also injures the rapidly dividing cells of the oral epithelium and cells of the connective tissue, causing a breakdown in the integrity of the oral mucosa. Oral mucositis often ensues, characterized by erythema, ulceration, and pseudomembrane formation. This condition can significantly complicate the treatment of the underlying malignancy, at times representing a dose- and volume-limiting toxicity, negatively impacting the efficacy of the treatment and greatly affecting the quality of life of the patient. Treatment-induced mucositis can also add significantly to the cost of the treatment.3 The pain from oral mucositis can be devastating, frequently requiring the use of narcotic analgesics for effective pain management and at times impairing oral intake to the extent that parenteral or enteral nutritional supplementation is necessary.4-7
The severity of mucositis is thought to be modulated by environmental factors, perhaps especially infection.8,9 Cultures show that patients with head and neck cancer receiving treatment with radiotherapy have colonization of the oropharyngeal region with gram-negative bacteria. Enterobacteriaceae and Acinetobacter species and possibly their endotoxins as well as viridans streptococci, staphylococci, and Enterococcus faecalis may have a significant role in the severity of radiation mucositis, especially perhaps in the development of pseudomembranes. Endotoxin, for example, is a potent inciter of inflammation.9-10 In contrast, yeast colonization has been shown not to be a significant determinant in the severity of treatment-induced mucositis.11
Early studies tested local administration of available antimicrobial agents, which were inconsistently effective in modifying the course and sequelae of oral mucositis.12-24 More recently, selective decontamination of the oral cavity with antimicrobial agents has demonstrated clinical benefit and a reduction in oral mucositis associated with radiation therapy.25-30 Four randomized trials have tested polymyxin E (2 mg), tobramycin (1.8 mg), and amphotericin B (10 mg) (APT) lozenges for the prevention of radiation mucositis in head and neck cancer patients. One of these trials has shown microbiologic efficacy but no clinical benefit,31 whereas the other three have shown clinical benefit to varying degrees.32-36 Requirements for an effective antimicrobial strategy include an agent with a sufficient spectrum of antimicrobial coverage, a low probability of developing resistant pathogens, and the ability to achieve and maintain an adequate reduction of oral flora. Taking into consideration the high estimated cost of the APT lozenge, the question arose regarding the viability of alternative antimicrobial lozenges that cover the same spectrum but with a significantly lower cost. We have carried out a pilot study with such a lozenge, which has shown that a less costly lozenge containing bacitracin, clotrimazole, and gentamicin (BCoG) with activity similar to the APT lozenge, can consistently suppress the gram-negative bacilli and yeast counts.37 The randomized trial reported here was conducted on the basis of the results of that pilot study with the aim of testing the clinical efficacy of the BCoG lozenge.
In addition to the above, almost all previous scoring methods used in these trials have been an amalgam of symptoms and physical appearance, which may not allow for reproducibility and consistency in assessing the outcome of interest. Moreover, nonvalidated measures detract from the validity of the results and comparability between studies. It was opportune that the newly validated mucositis scoring system, the Oral Mucosa Assessment Scale (OMAS),38 was available and used in this trial. The OMAS allows for a detailed assessment of mucositis in an objective and reproducible way.
PATIENTS AND METHODS
Patient Population
Patients included on the study had histologically confirmed nonmetastatic carcinoma of the oral cavity, pharynx (nasopharynx, oropharynx, or hypopharynx), or larynx; Eastern Cooperative Oncology Group performance status of 0, 1, 2, or 3; and were to receive radical or postoperative radiotherapy to a significant part of directly visible (two or more anatomic sites at risk) oral and/or pharyngeal mucosa. Patients were excluded if cytotoxic chemotherapy was to be given concurrently with the radiotherapy, or if they had received previous chemotherapy or radiation treatment to the upper airways or had a history of hypersensitive reaction to bacitracin, clotrimazole, or gentamicin. Consent was obtained according to local institutional and/or university human experimentation committee requirements. The patients were able (ie, sufficiently fluent) and willing to complete the quality-of-life questionnaires in either English or French and were accessible for treatment and follow-up.
Study Design
Enrolled patients were randomized to either arm 1, to receive radiotherapy to the oral cavity, pharynx, or larynx and BCoG lozenge (one lozenge qid, day 1 to end of radiotherapy); or arm 2, to receive radiotherapy to the oral cavity, pharynx, or larynx and placebo lozenge (one lozenge qid, day 1 to end of radiotherapy). Each BCoG lozenge contains the following: bacitracin 6 mg, clotrimazole 10 mg, and gentamicin 4 mg plus nonmedicinal ingredients. The BCoG lozenge is active against gram-positive cocci, gram-negative bacilli, and yeast microorganisms.39 Patients were stratified for center, radical versus postoperative radiotherapy, planned radiation target total dose of ≤ 55 Gy versus more than 55 Gy, and dose/fraction of ≤ 2 Gy versus more than 2 Gy.
A radiation oncologist or oral medicine practitioner carried out objective scoring of the mucosal reaction. The patients were seen weekly during radiation treatment, then 3 to 4 weeks after treatment, and again every 3 to 4 weeks until the resolution of mucositis. The OMAS requires grading of erythema as none, severe, or not severe; oral mucosal ulceration or pseudomembrane was graded as none, less than 1 cm, 1 to 3 cm, and more than 3 cm.38 The observer recorded the grade of erythema or ulceration/pseudomembrane in any of the following locations: upper and lower lips, right and left cheeks, right and left ventral and lateral tongue, floor of mouth, soft palate/fauces, and hard palate. The total mucositis score is determined on the basis of a sum of the at-risk or radiation-exposed sites. Oral pain was assessed using a visual analog scale for oral pain at rest and with function.
During the study, the following oral care was allowed: supportive care measures, baking soda mouth rinse, benzydamine rinse (Tantum; 3M, London, Ontario, Canada), Xylocaine viscous, acetaminophen/codeine, or morphine solution. Antibiotics and antifungals were only allowed as indicated by culture and sensitivity results. All patients were given pretreatment dietary counseling, reinforced weekly. Dietary supplements were available to all patients unable to consume adequate calories for maintenance (determined from record of weight change and history). A feeding tube was placed when and if needed, according to institutional protocol.
In terms of the specifications of radiation therapy, there was no real-time radiotherapy review in this study, but it was indicated that all planned fields must be treated every day with no more than one fraction per day, five fractions per week without any intended gaps and a planned target dose of ≥ 40 Gy.
Outcome Measures
The primary end point of the study was time to the development of objective mucosal end point (TTDOME), defined as the time from the first radiotherapy treatment day, when radiation was delivered to the oral mucosa in a standardized way and patients were given either BCoG lozenges or placebo lozenges, until the observer records grade 2 or 3 ulceration/pseudomembrane in any of the following locations: upper and lower lips, right and left checks, right and left ventral and lateral tongue, floor of mouth, soft palate/fauces, and hard palate.
The secondary end points of the study included the extent of severe mucositis score, worst-ever grade of ulceration/pseudomembrane, number of treatment days lost, changes in body weight, worst-ever grade of oral toxicity measured by patient diary, and general nonmucosal toxicity assessed by the investigator. The extent of severe mucositis score was defined by Sonis et al38 as the number of sites with either ulceration = 3 or erythema = 2. The final score for each patient was calculated over time as the average of the three highest values. Worst-ever grade of ulceration/pseudomembrane for one patient was calculated as the maximum of ulceration/pseudomembrane scores during the study. The total number of treatment days lost was recorded in the case report form as total numbers of days of interruption/delay when treatment was not given to a patient at the time specified in the protocol. Body weights of the patients were measured at baseline and at each clinical visit (during treatment and after treatment). The change of body weight was defined as the change from the baseline measurement.
Statistical Analysis
Planned sample size was 130 patients to allow the detection of a 10-day or larger difference in median TTDOME between the two arms at a two-sided 5% level test with 80% power. The sample size was increased from the target size of 130 to 138 to allow for potentially missing data and late exclusions.
The TTDOME was described by Kaplan-Meier curves and the stratified log-rank test was the primary method to compare TTDOME between the two treatment groups. The Wilcoxon rank sum test was used to compare the extent of severe mucositis score, distribution of worst-ever grades of ulceration/pseudomembrane, number of treatment day lost, and distribution of worst-ever grades of oral toxicity measured by patient diary. A nonparametric smoothing method, LOESS, was used to describe the time trend of body weights. A linear mixed model was used to test the interaction between treatment and time.
RESULTS
Patient Populations
One hundred thirty-eight patients were enrolled between September 1997 and September 1999 (69 on the BCoG arm and 69 on the placebo arm). One patient on the placebo arm was classified as ineligible and excluded from this analysis.
All the pretreatment characteristics of the patients (Table 1) and their tumors (Table 2) were balanced between the two arms of the study. Details of the radiotherapy treatments were similar between the two arms of the study (Table 3). The majority of patients (87% in total, with 86% on arm 1% and 88% on arm 2) were randomized onto the study within 12 weeks from the time of histologic diagnosis. Approximately 50% of the patients were randomized within 4 to 7 weeks.
Table 1. Patient Characteristics
Table 2. Tumor Characteristics
Table 3. Treatment Characteristics
TTDOME
Four patients (one on the BCoG arm and three on the placebo arm) did not have the assessment after the baseline, and the date of first radiotherapy treatment for one patient on the BCoG arm was missing. Because TTDOME was difficult to calculate for these five patients, they were excluded from this analysis.
The Kaplan-Meier curves for TTDOME are presented in Fig 1. There is no statistical difference between the two arms. The median TTDOME for patients taking BCoG was 3.61 weeks and 3.96 weeks for patients assigned to take placebo. At 3 weeks after the first radiotherapy treatment, 59% of patients on BCoG and 55% of patients on placebo did not have grade 2 or 3 ulceration/pseudomembrane on any of the mucosal surfaces under study. The P value of the (unadjusted) log-rank test was .61. After adjusting the stratification factors including center, the P value of stratified log-rank test was .12.
Fig 1. Time to development of objective mucosal end point by treatment arm.
Extent of Severe Mucositis Score
All 137 patients were included in the analysis of secondary end point of the extent of severe mucositis (69 on the BCoG arm and 68 on the placebo arm). The average severe mucositis scores (with SE) for patients treated by BCoG and placebo were, respectively, 2.59 (SE = 0.28) and 2.61 (SE = 0.33). The P value of the Wilcoxon rank sum test for the comparison of scores between two arms was .97. No statistically significant difference between the two arms in terms of the extent of severe mucositis score was seen.
Worst-Ever Grades of Ulceration/Pseudomembrane
Table 4 lists, for each arm, the numbers and percentages of patients whose worst-ever grades of ulceration/pseudomembrane are, respectively, 0, 1, 2, and 3. No statistical difference was seen between the two arms in terms of the distribution of worst-ever grades of ulceration/pseudomembrane (P = .54, Wilcoxon rank sum test).
Table 4. Distribution of Most Severe Grades of Ulceration/Pseudomembrane
Number of Treatment Days Lost
Among 68 patients on the BCoG arm, 20 (29%) experienced at least 1 treatment day of interruption/delay caused by treatment reaction. The corresponding number for 68 patients on the placebo arm was 16 (24%). This information is missing for three patients on the BCoG arm and two patients on the placebo arm. When only those patients with complete data sets are considered, the average number of treatment days lost was 0.57 for the 65 patients on the BCoG arm and 0.44 for the 66 patients on the placebo arm. The difference was not statistically significant (P = .48, Wilcoxon rank sum test).
Change of Body Weight
The plots of the average body weights of the patients on the two treatment arms as a function of the times since the patients were randomized onto the study are given in Fig 2. This figure indicates that patients on both arms lost weight after they received treatment. There is a statistically significant interaction in body weight loss between the treatment patients received and the time at which body weights were measured (P = .025). This means that the differences in body weight loss between the two arms are not the same at different time points. There is a trend, as time goes on, for the patients on the placebo arm to lose more weight than the patients on the BCoG arm. However, we failed to identify a time point at which there was a statistically significant difference in body weight loss between two arms.
Fig 2. Body weight by arm.
Oral Toxicities Measured by Patient Diary
Table 5 lists, for each arm, the numbers and percentages of patients whose worst-ever answers are, respectively, 0 (not at all), 1 (a little), 2 (quite a bit), and 3 (very much) to each of the questions asked in the patient diary (some questions may not be applicable to all the patients; therefore, only those who answered the questions were included in the analysis). The P values of the Wilcoxon rank sum tests for the comparison between the two treatment arms are also listed in Table 5. More patients taking BCoG had significant change in taste (P = .04).
Table 5. Worst-Ever Grades of Oral Toxicities Recorded in Patient Diary
Other Nonmucosal Toxicities
Detailed nonmucosal systematic acute and delayed toxicities were carefully documented according to the National Cancer Institute Clinical Trials Group, Expanded Common Toxicity Criteria. These included cardiovascular, gastrointestinal, pulmonary, neurologic, and skin side effects in addition to infection and flu-like illnesses. Toxicities were as expected in this cohort of patients treated with conventional radiotherapy. One patient on the BCoG arm died of secondary malignancy. Other toxicities were comparable between the two arms of the study.
DISCUSSION
Oral mucositis has significant clinical implications in patients receiving radiation therapy for treatment of head and neck cancer. It may cause severe pain and dysfunction that interfere with swallowing and speech and lead to serious consequences such as weight loss. More importantly, radiation-associated mucositis can have significant implications on tumor control or cure, if treatment has to be interrupted to allow for healing or the dose and volume have to be reduced. Indeed, attempts at reducing severe mucositis by decreasing the total dose of radiation, the volume of treated mucosa,40 and/or splitting the course of radiation to allow time for healing has proven detrimental to the control or cure of the cancer in some cases.41
In this phase III trial, we sought to achieve two objectives: to determine whether a microbiologically efficacious and less expensive alternative to the more costly APT lozenge could reduce the incidence of mucositis and its consequences in patients treated with conventional radiotherapy; and to use a newly validated mucositis scoring system38 in the context of a large randomized trial as a means of demonstrating the utility of a standardized approach in the assessment of this outcome. Our study included a fairly homogenous population of patients who were treated with conventional radiotherapy alone, with the majority treated to radical doses above 55 Gy. The dose per fraction was around 2 Gy in the majority of patients. Most patients were suffering from carcinoma of the oral cavity/oropharynx. This ensured that we minimized the variables that could have interfered with the main question tested in the study.
There are several possible explanations why our study showed no benefit when others that tested lozenges with a similar antimicrobial spectra (Table 6) were positive: (1) the lozenges did not achieve the same degree of antimicrobial control as those in previous studies; (2) the outcome measure was insensitive; (3) the radiation protocols used were different; and (4) the standard of care in oral hygiene and social habits were different. The first two of these explanations do not seem particularly plausible. Separate investigations indicated that the lozenges did have the expected antimicrobial activity.37 The outcome measures were derived from a validated scoring system.38 The populations studied in the trials were similar in the tumor variables. Conversely, variations did exist in the radiation protocols and the allowable basic oral hygiene. Finally, although patient self-report of symptoms proved to be a more sensitive end point in one trial, other positive trials used the observations of health care personnel.
Table 6. Published Studies of Oral Topical Antimicrobial Lozenges in Radiation Therapy of Head and Neck Cancer
We conclude, therefore, that the results of our trial are both valid and representative and show that the prophylactic use of antimicrobial lozenges is unlikely to play an important role in controlling the mucositis caused by conventional-dose radiotherapy. Other approaches to this devastating symptom are clearly needed, particularly as the treatment becomes more aggressive, leading to higher rates of acute toxicities.42-60 They should be examined in carefully designed studies using an objective validated scoring instrument. In the meantime, new technology in radiation treatment such as intensity-modulated radiotherapy might enable further reduction of the treated mucosal volume without compromising the chance of tumor control.
Acknowledgments
Supported and administered by the Clinical Trials Group, National Cancer Institute of Canada, Kingston, Ontario, Canada.
ACKNOWLEDGMENT
We thank the following investigators for contributing patients to the study: D. Donath, R. Choo, J. Kim, M. Anthes, V. Malik, E. Sham, B. Sheehan, and C. Leong.
- Received May 6, 2002.
- Accepted June 19, 2002.
