Erratum to “Clinical evaluation of the intraoral fluoride
releasing system in radiation-induced xerostomic subjects. Part 2: Phase I study”
,
Mark S. Chambersa, d, 
, 
, Terence J. Fleminga, Béla B. Totha, James C. Lemona, Timothy E. Cravenb, Otis J. Bouwsmac, Adam S. Gardend, Mark A. Espelandb, Harris J. Keenea, Jack W. Martina and Tibor Sipose
aDepartment of Head and Neck Surgery, The University of Texas, M. D. Anderson Cancer Center (MDACC), Section of Oncologic Dentistry, Box 441, 1515 Holcombe Boulevard, Houston, Texas 77030-4009, USA
bDepartment of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, NC, USA
cPeriodontist, Rancho Viejo, Texas, Consultant, Digestive Care Inc., Bethlehem, PA, USA
dDepartment of Radiation Oncology, MDACC, USA
eDigestive Care Inc., Bethlehem, PA, USA
Available online 8 December 2006.
| Refers to: |  |
Clinical evaluation of the intraoral fluoride releasing system in radiation-induced xerostomic subjects. Part 2: Phase I study Oral Oncology, Volume 42, Issue 9, October 2006, Pages 946-953 Mark S. Chambers, James R. Mellberg, Harris J. Keene, Otis J. Bouwsma, Adam S. Garden, Tibor Sipos, Terence J. Fleming  PDF (124 K)
|
| Refers to: | |
Clinical evaluation of the intraoral fluoride releasing system in radiation-induced xerostomic subjects. Part 2: Phase I study Oral Oncology, Volume 42, Issue 9, October 2006, Pages 946-953 Mark S. Chambers, James R. Mellberg, Harris J. Keene, Otis J. Bouwsma, Adam S. Garden, Tibor Sipos, Terence J. Fleming PDF (124 K)
|
Summary
Radiation-induced xerostomia can result in the rapid onset and progression of dental caries in head and neck cancer patients. Topically applied 
fluorides have been successfully used to inhibit the formation of dental caries in this population. However, because intensive daily self-application is required, compliance is an issue. The intraoral fluoride-releasing system (IFRS) containing a sodium fluoride core is a newly developed, sustained-release, passive drug delivery system that does not require patient involvement except for periodic replacement, thus reducing the effect of patient compliance on its effectiveness in dental caries prevention. Twenty-two head and neck cancer patients from U. T. M. D. Anderson Cancer Center, with radiation-induced xerostomia, were entered into a pilot study to contrast the daily home use of a 0.4% stannous fluoride-gel-containing tray (control group) to IFRS (study group) with respect to tolerability and adherence, and to obtain information on relative caries preventive efficacy. Participants were stratified on the basis of radiation exposure and randomly assigned to treatment with either IFRS or stannous fluoride gel. Patients in both groups were fitted with two IFRS retainers and also were instructed to use a 1100-ppm fluoride conventional sodium fluoride dentifrice twice daily. The study was conducted as a single-blinded, parallel-cell trial. Pre-existing carious lesions were restored prior to the beginning of the study. The efficacy variable was determined by the mean number of new or recurrent decayed surfaces. Patients were examined for caries 4, 8, 12, 24, 36, and 48 weeks after initiation of treatment. Reports of adverse reactions were based on information volunteered by patients and that were elicited during interviews. At baseline, the resting and stimulated salivary flow rates (g/5 min) were significantly greater in the control group than in the study group (p < 0.05). Patients in the control group had received significantly more radiation than those in the test group (68 Gy vs. 60 Gy; p = 0.047). No marked differences in follow-up new and recurrent caries were found between the stannous fluoride gel control and IFRS groups during the study period. The rate of new or recurrent carious lesions in the group treated with the fluoride gel was slightly lower than in the IFRS group, based on carious lesions at the baseline examination (Poisson mean number of new or recurrent carious lesions for the control group = 0.55 per year vs. 0.83 per year for the study group, p = 0.705; odds ratio of the occurrence of any new or recurrent caries during follow-up for control group vs. the study group = 0.80; p = 0.781). This pilot study revealed that the IFRS was well-tolerated and safe in this study population associated with minimal complications during the duration of this study and was comparable in efficacy to a SnF2 gel in preventing caries development. The IFRS provided similar rates of control for caries formation to a fluoride-gel-containing tray. The IFRS is designed to release a daily dose of 0.12 mg of sodium fluoride, which can be evenly distributed throughout the oral cavity for a single application of 4 months. It would be more convenient than the daily home application of a tray of 0.4% stannous fluoride or 1.1% sodium fluoride gel, and avoids the problem of variable patient compliance.
Keywords: Fluoride; Head and neck radiation; Xerostomia; Intraoral fluoride releasing system
Article Outline
Introduction
Part 1 of this series reviewed the efficacy of fluoride in preventing or reducing dental caries in human populations. This part will describe the clinical evaluation of a fluoride releasing system. The purpose of the present study was to compare the efficacy of daily use of 0.4% stannous fluoride-gel in custom-made carriers (control group) to that of the intraoral fluoride releasing system (IFRS study group). Both groups were also instructed to supplement their treatment with a sodium fluoride (1100 ppm fluoride) dentifrice, to be used twice daily. From among the several groups that are considered to be at high risk for dental caries development, the one selected for this study was head and neck cancer patients whose radiation treatment significantly reduced salivary flow.
Materials and methods
This pilot study was conducted among a subgroup of patients being treated in the Department of Head and Neck Surgery Section of Oncologic Dentistry and Prosthodontics at The University of Texas M. D. Anderson Cancer Center (MDACC). Approximately 3 months earlier, the patients had undergone radiation therapy to the head and neck region that involved the major salivary glands and had been placed on the conventional MDACC post-radiation 2-part intensive fluoride preventive home care protocol. The first part consisted of the twice daily use of an American Dental Association-approved sodium fluoride dentifrice with 1100 ppm fluoride (Colgate Regular; Colgate-Palmolive Co., New York, NY) applied with Colgate TotalTM toothbrushes (Colgate-Palmolive Co., New York, NY). The second part consisted of once daily 10-min fluoride gel application (Gel-Kam™, 0.4% stannous fluoride, 970 ppm fluoride; Colgate-Palmolive Co., New York, NY) in custom-made fluoride carriers. This treatment was continued until the immediate effects of the radiation subsided (i.e., mucositis, dermatitis, dysgeusia), about 3 months post-radiation therapy.
The study protocol was reviewed and approved by the Institutional Review Board of MDACC. All patients signed the consent form after the study purpose, procedures, risks, and benefits were explained to them. Twenty-two patients (six females and 16 males; 42–80 years of age, mean age 55.7 years) volunteered to participate in the study. This sample size was expected to provide adequate experience to contrast the tolerability and safety of the two treatments and to provide information on the relative treatment effects for use in designing a more definitive trial. The patients had to be at least 18 years of age and have a minimum of 10 natural teeth, exclusive of third molars, and at least two opposite bondable maxillary posterior teeth in the vicinity of the right and left Stensen’s (parotid) ducts. Patients had to have undergone radiation therapy totaling a minimum 40 Gy at least 3 months prior to study entry with radiation volumes including at least one submandibular or one parotid gland. Unstimulated salivary flow in these patients had to be less than 0.30 g/min. Patients had to be currently using the fluoride tray regimen and fluoride dentifrice as described above and could have no unrestored carious lesions at the beginning of the study. Those patients who had carious lesions at the time of screening underwent caries excavation and conventional restorative therapy prior to the study.
Sialometry was performed to capture unstimulated and stimulated salivary flow. Unstimulated whole saliva is a measure of total basal (resting) saliva, including major and minor gland secretions. Stimulated saliva is a measure predominantly of parotid gland secretions with less submandibular and minor salivary gland contributions. Resting and stimulated saliva flow tests were conducted 5 min apart at baseline and at each scheduled visit. The weight of the saliva sample was measured to an accuracy of 0.0001 g using a calibrated analytical balance. Unstimulated and stimulated whole saliva was collected in preweighed tubes and flow rates (g/5 min) were calculated after reweighing the tubes. A digital analytical balance was used for all weighings (TR Series; Denver Instruments, Arba, Colo). Subjects were initially asked to swallow any residual saliva and then to expectorate every minute for a 5 min period. A 2% citric acid solution applied every 30 s was used as the stimulant in the stimulated sialometry and a 2-min rinsing preceded the salivary collection.
Patients were excluded if they were receiving total body irradiation for bone marrow transplants; if they had Hodgkin’s disease or non-Hodgkin’s lymphoma with mantle radiation treatment; or if they had uncontrolled, clinically significant renal or cardiovascular disease. In addition, patients with brain cancer, osteomalacia, Crohn’s disease, any neurologic disease (i.e., Parkinson’s disease), gastric ulcers, taking investigational drugs, or allergy to fluoride were excluded. Women who were pregnant, trying to become pregnant, or breast-feeding were excluded as well. Patients with a known metastasis or second primary tumor in the head and neck region who were likely to receive additional radiation therapy were also excluded, as was any patient who indicated an unwillingness to comply with the protocol and treatment regimen.
Clinical and IFRS procedures
Following the 3-month post-radiation therapy waiting period, subjects during the first visit had the placement of the IFRS devices. Both the control (fluoride-gel-containing tray) and study group (IFRS) had the IFRS retainers placed onto the identifiable bilateral maxillary molars. The control group had the IFRS retainers placed with placebo tablets to keep the evaluators blinded to the treatment arm. The IFRS retainers were bonded to the buccal surfaces of the bilateral maxillary molars using Fuji II duet luting agent (GC America, Alsip, IL) and following the manufacturer’s instructions. The IFRS consists of two components: (1) a tablet for sustained-release of fluoride, hereafter referred to as an intraoral fluoride-releasing device (IFRD) and (2) a retainer that holds and protects the IFRD while in the mouth. The IFRD is a membrane-controlled reservoir-type device (half-football shaped) that consists of an inner core and an outer coating. The inner core contains a specific amount of powdered sodium fluoride (active ingredient) dispersed in a 50:50 mole ratio of acrylate copolymer composed of hydroxyethyl methacrylate (HEMA) and methyl methacrylate (MMA). The core is surrounded by a membrane coating, a pure film of a 30:70 mole ratio of HEMA/MMA copolymer, which controls the rate of fluoride released from the device. The paired IFRDs (active and placebo) were replaced every 3 months. Retrieved tablets were randomly analyzed for remaining sodium fluoride.
The study was conducted as a single blind, parallel-cell design. Following admission, participants were stratified on the basis of radiation exposure (complete or partial; complete if it included bilateral parotid and submandibular glands, and partial if the field of radiation includes bilateral submandibular glands only or unilateral parotid and submandibular glands) and randomly assigned to one of two treatment cells: the IFRS treatment or fluoride gel control group. Separate variable block length randomization schedules were prepared for each radiation exposure strata. Subjects were entered sequentially, and randomly assigned into one of two cells: 10 to IFRS treatment (study group) and 12–0.4% stannous fluoride gel tray treatment (control group). All subjects were given verbal and written oral care instructions advising them to continue using the twice daily 1100 ppm fluoride toothpaste protocol and to brush after each meal. The control group continued to apply the once daily 0.4% stannous fluoride on their teeth before bedrest using a gel-in-carrier application for 10 min and remaining n.p.o. for 30 min. The efficacy variable was the mean number of new or recurrent decayed surfaces. The visual-tactile qualitative and quantitative criteria of Radike1 were used to determine coronal caries. The criteria used to define root caries and quantitate severity of root caries were those of Banting et al.2 The root caries index (RCI) of Katz3 was used to quantify root caries. Examinations were scheduled at 4, 8, 12, 24, 36, and 48 weeks after study initiation. The paired IFRDs were replaced every 12 weeks and subjects had reinforcement of oral care instructions at each study visit.
Subject compliance to the control treatment regimen was assessed by the following mechanisms: (a) subjects were given a home diary card to record the time and date of once-daily gel application; and (b) the quantity of gel used by each subject was monitored by allocating tubes of gel to the subjects at each visit and maintaining a record of returned tubes and the volume of gel used per subject.
Safety examinations
Each subject underwent hematologic and renal tests at baseline and at 3, 6, and 9 months. Hematology tests included hemoglobin, hematocrit, platelets, white blood cell count, and differential. Blood chemistry panels included calcium, inorganic phosphorus, sodium, potassium, chloride, AST, total bilirubin, ALT, alkaline phosphates, cholesterol, total protein, albumin, BUN, creatinine, uric acid, glucose, and CO2. Urinalysis included specific gravity, pH, WBC, RBC, CASTS, protein, glucose, and ketones. Additionally, a safety evaluation was performed at each clinic visit by a subjective evaluation of patient diaries, systems review, oral cavity examination, and an attitudes/symptoms quality of life questionnaire.
Statistical analysis
The counts of coronal and root caries were assumed to follow a Poisson probability distribution (a common assumption for discrete data), which appeared appropriate based on histograms. As specified in the study protocol, statistical analysis of the primary outcome measure (occurrence of coronal and root caries) was performed using a standard method for analyzing repeated counts over time: generalized estimating equations.4 and 5 This approach, while similar in many ways to repeated measures of analysis of variance for normally-distributed continuous data, is more appropriate for counts. The secondary outcome (time to failure) was analyzed using the proportional hazards regression model, in which patients who did not develop three or more new or recurrent caries were censored. All statistical tests were performed at the 5%, two-sided significance level using the intent-to-treat principle of analysis. When three new or recurrent carious lesions were observed, a subject was considered to have experienced an “event”. Subjects with two or fewer carious lesions observed during the follow-up period were considered “censored”. Treatment-arm differences in the time-to-occurrence of three new or recurrent caries were evaluated using a Cox proportional hazards regression model, with treatment as the single covariate. Time to the event of interest was calculated using the midpoint of the follow-up interval in which the event occurred as the observed event time. Censoring times were set at 48 weeks for participants who completed the study as planned and the time of the last clinic evaluation for subjects lost to follow-up.
The RCI was calculated at baseline and at each follow-up evaluation (4, 8, 12, 24, 36, and 48 weeks after study initiation) using the following formula:
Other end points evaluated were response to the 12-item attitudes/symptoms questionnaire and the simple average of the 12 responses. All responses were recorded using the following scale (assigned score in parentheses): “much worse” (−2), “worse” (−1), “no change” (0), “better” (+1), or “much better” (+2). Thus, positive scores are associated with attitude/symptom improvement, negative scores with attitude/symptom worsening, and zero scores with no change. Each of these end points was analyzed via ML-ANOVA models with pretreatment score as a covariate.
Initially, blood chemistries were required by the IRB as a method of monitoring possible fluoride toxicity. Blood chemistries were done at baseline and at the 4-, 8-, 12- and 24-week examinations. No changes were observed that would have indicated a toxicity problem; therefore, the IRB approved the request to discontinue the blood chemistries.
Results
Patient characteristics
A comparison of the pre-trial characteristics of the control and IFRS groups is shown in Table 1. There was a greater number of men randomly assigned to the control group compared with the IFRS group (91% vs. 50%; p = .056). At baseline, both unstimulated and stimulated saliva flow rates were greater in the control group than in the study group (0.697 vs. 0.285 g/5 min; p = 0.003 and 0.921 vs. 0.577; p = 0.126). The subjects in the control group received significantly more radiation than those in the study group (68 Gy vs. 60 Gy, respectively; p = 0.047). The mean number of teeth per subject was not different between the control and study groups (24.3 vs. 23.0, respectively; p = 0.374). The number of coronal and root caries and the RCI (mean) were numerically higher in the control group, but the intergroup differences were not statistically significant (1.67 vs. 0.200, respectively; p = 0.226: 1.88 vs. 0.370, respectively; p = 0.127). As the data in Table 1 indicate, the subjects in both groups were aware of the importance of fluoride and the importance of daily oral care. Compliance, as determined by subject questionnaires and verbal communication, was considered good as has been demonstrated by similar patients in this institution.
Pre-trial descriptive statistics for screening factors (means ± S.D.)
| Factor | Fluoride gel (n = 12) | IFRS (n = 10) | p-value |
|---|---|---|---|
| Age (years) | 55.4 ± 10.7 | 56.0 ± 10.5 | 0.899 |
| Height (cm) | 179.1 ± 3.6 | 176.9 ± 8.2 | 0.454 |
| Weight (kg) | 80.2 ± 11.5 | 72.0 ± 11.8 | 0.115 |
| Unstimulated salivary flow (g/5 min) | 0.70 ± 0.3 | 0.30 ± 0.3 | 0.003 |
| Stimulated salivary flow (g/5 min) | 0.92 ± 0.42 | 0.60 ± 0.6 | 0.126 |
| Radiation total dose (cGy) | 6825 ± 893 | 6044 ± 823 | 0.047 |
| Q: “How dry is your mouth now?”a | 2.33 ± 0.89 | 2.10+ ± 1.29 | 0.621 |
| Q: “How important is fluoride?”a | 4.67 ± 0.49 | 4.30 ± 0.82 | 0.211 |
| Q: “How important is daily oral care?”a | 4.75 ± 0.62 | 4.40 ± 1.27 | 0.440 |
| Number of teeth | 24.3 ± 3.6 | 23.0 ± 3.2 | 0.374 |
| Number of coronal and root caries lesionsa | 1.67 ± 3.92 | 0.20 ± 0.632 | 0.226 |
| Root caries index (lesions)a | 1.88 ± 2.82 | 0.37 ± 1.11 | 0.127 |
| Male gender | 11 (92%) | 5 (50%) | 0.056 |
| Caucasian | 11 (92%) | 9 (90%) | 1.000 |
| Partial radiation stratum | 2 (17%) | 0 (0%) | 0.481 |
At the baseline evaluation, participants were asked to rate three symptoms or attitudes related to their dental care, compared to when they felt “normal”. The participant responded to each of the following questions by rating it on a 5-point scale: (1) “How dry is your mouth now?”, rated as 1 = severely dry, 2 = very dry, 3 = dry, 4 = mildly dry, 5 = not dry; (2) How important is the use of fluoride gel in your daily oral care now?” and (3) “How important is your daily oral care now?”, each rated as 1 = not important, 2 = slightly important, 3 = important, 4 = very important, 5 = extremely important. Mean scores for each question are shown in Table 1 by treatment arm assignment. Average scores for each of the three questions did not differ significantly between treatment groups.
Clinical examinations
Analyses included participants who were evaluated at 4, 8, 12, 24, 36, and 48 weeks of follow-up. Some participants were followed for a period of time after the planned study closeout at 48 weeks; however, only the data gathered during the 48-week active study period were included in the analyses presented here. Table 1 presents descriptive statistics at screening by treatment arm (see Table 2).
Adverse events (n = 14)
| Treatment arm | N | Adverse event |
|---|---|---|
| IFRS | 1 | Constipation |
| 1 | Dull chest pain | |
| 3 | Caries | |
| Fluoride gel | 2 | Buccal mucosa irritation |
| 7 | Caries |
The 22 subjects, randomly placed either into the control group (fluoride gel; 11 men, 1 woman) or the test group (IFRS; 5 men, 5 women), were found to be similar at baseline. Most of the differences could be attributed to the greater number of men randomly assigned to the control group (11 men, 91%) compared with the study group (5 men, 50%). The average age in years for the control and study groups was not significantly different (55.4 years vs. 56.0 respectively; p = 0.899). Likewise, the heights of the subjects were similar for the control and test groups (179.1 cm vs. 176.9, respectively; p = 0.454). The larger weight (in kilograms) of patients in the control group compared with those in the study group might be attributable to the greater percentage of men in the control group (80.2% vs. 72.0% respectively; p = 0.115).
Primary outcome-incidence of new/recurrent caries
A basic assumption of the proposed analysis of the primary outcome, incidence of new or recurrent coronal or root caries during the 48-week follow-up period, is that the incidence or recurrence of caries follows a Poisson distribution over any fixed observation period. Upon evaluation of the data collected, the number of patients who experienced new or recurrent caries was small (six total, three in the fluoride gel control group and three in the IFRS group). Among individual subjects who had caries, the number of carious lesions observed was quite large in some cases (range, 1–17). Table 3 shows the distribution of the number of carious lesions and the number of subjects experiencing those lesions, by treatment and week of study.
Distribution of the number of carious lesions and the number of patients experiencing lesions by treatment and week of study
| Treatment | Study week
|
|||||
|---|---|---|---|---|---|---|
| Screeninga | 12 | 24 | 36 | 48 | ||
| Fluoride gel group | ||||||
| Total number of patients | 12 | 12 | 11 | 11 | 11 | |
Number of patients with  1 new or recurrent lesions |
2 | 2 | 2 | 3 | 2 | |
| Total number of carious lesions | 16 | 16 | 16 | 23 | 17 | |
| Rate of occurrence of carious lesions per person in interval | 1.33 | 1.33 | 1.46 | 2.09 | 1.55 | |
| IFRS group | ||||||
| Total number of patients | 10 | 10 | 9 | 9 | 9 | |
| Number of patients with 1 new or recurrent lesions |
1 | 1 | 2 | 1 | 1 | |
| Total number of carious lesions | 1 | 1 | 4 | 3 | 5 | |
| Rate of occurrence of carious lesions per person in interval | 0.10 | 0.03 | 0.44 | 0.36 | 0.56 | |
The presence of carious lesions at the initial screening period was the most important factor related to the development of new or recurrent caries during the follow-up treatment period; therefore, all models for treatment efficacy included the baseline (prestudy number of carious lesions) as a covariate. Thus, the hypothesis of interest is rate of new or recurrent lesions during the follow-up period on the basis of the disease status at randomization. Two methods were utilized to evaluate this hypothesis: (1) A GEE Poisson regression model used to evaluate efficacy with respect to the number of new or recurrent lesions observed during the on-treatment follow-up; (2) A GEE logistic regression model used to evaluate efficacy with respect to the occurrence of any new or recurrent lesions during the on-treatment follow-up. For both models, treatment × (baseline outcome) and treatment × (follow-up month) interactions were evaluated and found to be nonsignificant. Thus, models for efficacy included main effects for treatment, baseline outcome, and a month of follow-up. Table 4 summarizes the results of the two models for treatment efficacy.
Results of GEE models for treatment efficacy
| Models | Link | Treatment effects
|
||||
|---|---|---|---|---|---|---|
| Beta | S.E. | Measure of association | Value (95% CI) | p-values | ||
| GEE Poisson regression | Log | −0.4148 | 1.1063 | μ (fluoride gel)/IFRS | 0.66 (0.08, 5.77) | 0.705 |
| GEE logistic regression | Logistic | −0.2172 | 0.7827 | OR (fluoride gel/IFRS) | 0.80 (0.17, 3.73) | 0.781 |
μ = mean; OR = odds ratio; S.E. = standard error.
No significant differences in follow-up new or recurrent caries were found between the fluoride gel and IFRS groups when tested with either the GEE Poisson regression model or the GEE logistic regression model. The rate of new or recurrent caries for patients in the fluoride gel group was 1.55 per year compared with 0.56 per year in the IFRS group (p = 0.705); or of any new or recurrent caries during follow-up for fluoride gel group versus the IFRS group (0.80; p = 0.781). There is no evidence of a differential effect on the rate or occurrence of new or recurrent caries between the two treatment groups.
Secondary outcomes-root caries index (RCI) and time to three new/recurrent caries
Table 5 summarizes the baseline and follow-up RCI values by treatment arm and week of study follow-up. Subjects with (RD + RF + RN) equal to 0 were excluded from analysis. Treatment by time interactions were evaluated and found to be non-significant; thus, the test for treatment efficacy used the on-treatment RCI average over all follow-up exams. The estimated average on-treatment RCI (±S.E.), after controlling for pre-treatment values, were 0.95 ± 0.19 in the fluoride gel group and 1.1 l ± 0.21 in the IFRS group. These differences in mean RCI were not statistically significant (p-value = 0.600). Among the 12 fluoride gel subjects, two events were observed (both at 2 weeks of follow-up). Among the 10 IFRS subjects, one event was observed (at 18 weeks of follow-up). All other subjects were censored. These data yielded an estimated hazard ratio of 1.77 for the fluoride
