Oral mucositis in patients undergoing radiation treatment for head and neck carcinoma

Radiation-induced oral mucositis (OM) is common among patients with head and neck carcinoma (HNC), and is the single most debilitating side effect of radiation treatment.1 Acute severe OM is associated with significant discomfort and impairment of the patient's ability to eat and swallow. OM also is related to breaks in treatment, the placement of feeding tubes, and hospitalization.2-4

Despite its frequency and clinical impact, risk factors for OM in HNC patients have not been well defined to date. The incidence of mucosal injury has been reported to vary with radiation intensity and schedule and chemotherapy agent, dose, and schedule.2, 5-25 Hyperfractionated regimens and the use of conditioning and concomitant chemotherapy in particular increase the risk, severity, and duration of mucosal injury.2, 25 Risk factors related to patient characteristics are less clear. Age, gender, body mass index, use of alcohol and tobacco, preexisting dental disease, advanced disease and tumor stage, low leukocyte count, and changes in salivary flow and oral bacterial flora have been identified as possible risk factors for OM, but data supporting the significance of each of these measures are limited and uneven.

Although the standardized grading of mucositis severity with instruments such as the World Health Organization (WHO) or National Cancer Institute–Common Toxicity Criteria (NCI-CTC) scales is the norm in some practices and in randomized controlled clinical trials,2 the evaluation of OM based on subjective observation captured by narrative documentation is not unusual in other settings. The objective of the current study was to examine risk factors for OM in HNC patients receiving radiation treatment in a broad range of practice settings, and to explore the correlation between OM and adverse clinical outcomes that may be associated with mucosal injury.

MATERIALS AND METHODS

The current study was based on a nationwide cross-sectional survey of U.S. medical and radiation oncologists. Eligible oncologists had to be actively treating 3 or more HNC patients (age ≥ 18 yrs) who completed a course of radiation therapy within the preceding 12 months; 50% or more of their professional time also had to be involved in direct patient care. A total of 500 oncologists were invited (via telephone interview) to participate in the study and were sent case report forms. Each oncologist was asked to collect data from chart review for up to six patients. A total of 154 oncologists participated in the study.

Data were collected regarding patient characteristics, including age, gender, body weight, prior and/or current alcohol and tobacco use, primary tumor location, status (new vs. recurrent disease), stage of disease (AJCC Stages II, III or IV), and type of insurance. Information also was collected regarding the type of radiation therapy received (standard, hyperfractionation, standard plus hyperfractionation, or accelerated) during the last completed course, the dose given (given in centigrays [cGy], per day and cumulative), and whether the patient also received chemotherapy (before, during, or after radiation therapy or never). The highest grade of OM that patients experienced during radiation therapy (none, mild, moderate, or severe) was noted based on investigator judgment. Data also were collected regarding clinical outcomes (and the reasons therefore) that may be associated with mucosal injury, including delays and breaks in radiation therapy or chemotherapy, chemotherapy dose reductions, the use of feeding tubes or total parenteral nutrition (TPN), the placement of indwelling intravenous (i.v.) catheters, and hospitalization.

The correlation between selected patient characteristics and the highest recorded grade of OM was investigated first. Risk factors were examined alternatively for any OM versus no OM, moderate or severe OM versus no OM, and severe OM versus no OM. Chi-square and Fisher exact tests, one-way analysis of variance (ANOVA), and Kruskal–Wallis statistics, as appropriate, were used to assess differences in these characteristics in relation to the highest recorded grade of OM. A P value of < 0.05 was used to assess statistical significance. Polytomous logistic regression26 also was used to estimate the risks of any, moderate or severe, and severe OM, respectively (all vs. no OM); adjusted odds ratios (ORs) and corresponding 95% confidence intervals (95% CIs) were calculated. Variables in these models included patient age, gender, and body weight; alcohol and tobacco use; tumor location, status, and stage; radiation type and mean cumulative dose; and whether the patient received concomitant chemotherapy.

We next examined the correlation between the highest recorded grade of OM and selected clinical outcomes that may be associated with mucosal injury, including breaks and delays in radiation treatment and chemotherapy, chemotherapy dose reductions, the use of feeding tubes or TPN (for reasons other than prophylaxis), the placement of an indwelling i.v. catheter, and hospitalization.

Descriptive statistics were used to examine the distribution of these clinical outcomes for patients with no OM and mild, moderate, and severe OM, respectively. Chi-square and Fisher exact tests were used, as appropriate, to assess statistical significance, using a P value of < 0.05. The independent contribution of “any OM” to each of the outcomes of interest was assessed using dichotomous logistic regression, controlling for patient age, gender, and body weight; alcohol and tobacco use; tumor location, status, and stage; insurance type; radiation type and mean cumulative dose; and the presence of concomitant chemotherapy with radiation therapy. Separate regression models were estimated for each outcome. Adjusted ORs and 95% CIs subsequently were calculated for any OM and each of the outcomes of interest.

RESULTS

Data were obtained for 450 HNC patients treated by 154 oncologists. Approximately 62% of the practices were office based whereas 38% were based in (principally community) hospitals. Approximately 57% of participating oncologists did not routinely use standardized grading scales for OM. OM assessments were available for 437 study subjects.

Approximately 80% of the patients were male (Table 1). The mean (± the standard deviation [SD]) age was 61.3 years (± 12.3 yrs), and the mean body weight was 72.7 kg (± 18.1 kg). Approximately 80% of patients were either former or current alcohol or tobacco users. Primary tumor locations were the oropharynx (26.4%), larynx (26.4%), oral cavity including the lip (24.4%), hypopharynx (13.6%), and nasopharynx (9.1%). Most tumors were new and classified as Stage III (38.1%) or Stage IV (38.8%). The majority of patients (83%) received standard radiation therapy; the mean (± SD) cumulative dose was 6258 cGy (±1158 cGy). Approximately one-third of the patients received concomitant chemotherapy during radiation treatment.

The majority of patients (83%) developed OM (mild in 19%, moderate in 35%, and severe in 28% of patients). Table 2 summarizes the distribution of selected patient and treatment characteristics by the highest reported severity of OM. Patients with OM were on average younger (P < 0.01) and more likely to receive a higher cumulative radiation dose (P < 0.0001) compared with those without OM. Approximately 76% of patients receiving hyperfractionation regimens and 65% of those receiving standard therapy experienced moderate or severe OM; among those patients receiving combined standard and hyperfractionation radiation therapy, this number was 58%. Among the seven patients receiving accelerated regimens, none experienced moderate or severe OM. The severity of OM did not appear to vary significantly with respect to patient gender or body weight, prior or current alcohol or tobacco use, tumor stage, or primary insurance.

Nasopharyngeal or oropharyngeal tumor location, a cumulative radiation dosage > 5000 cGy, and the receipt of concomitant chemotherapy were found to increase the risk of OM in logistic regression models that controlled for patient age, gender, body weight, tumor stage, former or current alcohol or tobacco use, and type of radiation therapy (Table 3). The risk was lower for patients with advanced age (≥ 80 yrs). The risk of severe OM was found to be greatest for patients with nasopharyngeal tumors (adjusted OR of 10.1; 95% CI, 2.1–49.9), mean cumulative radiation dosages > 6500 cGy (OR of 10.4; 95% CI, 2.9–37.1), and concomitant chemotherapy (OR of 3.3; 95% CI, 1.4–8.0). Tumor location in the oral cavity (including the lip area) also was associated with a high risk of severe OM (OR of 6.2; 95% CI, 2.3–16.8). Patients with traditional indemnity insurance plans also were more likely to experience OM.

Unplanned breaks/delays in radiation therapy and chemotherapy were more common among patients with OM (P < 0.001 for both), and increased in frequency with the highest recorded grade of this complication (Table 4). Unplanned breaks/delays in radiation therapy due to “severe mucous membrane toxicity” were reported to occur among 2.4% of patients with mild OM, 15.8% of those with moderate OM, and 46.8% of those with severe OM. Corresponding estimates for unplanned breaks or delays in chemotherapy were 0.0%, 6.6%, and 25.0%, respectively. Patients with OM also were more likely to have had reductions in their chemotherapy dose (P = 0.001), the placement of feeding tubes or receipt of TPN (for reasons other than prophylaxis) (P = 0.009), and indwelling i.v. lines (P = 0.013). Approximately 33% of patients with severe OM were hospitalized compared with 16% of those with moderate OM, 21% of those with mild OM, and 11% of those with no reported OM (P < 0.001).

When patients with any OM were compared with those without OM on a multivariate analysis, the former group were found to be approximately 4-fold more likely to have had unplanned breaks in radiation therapy (adjusted OR of 3.8; 95% CI, 1.7–8.5) and more than 3 times as likely to have been hospitalized (OR of 3.5; 95% CI, 1.3–9.5) (Table 5). They also were 3.4 times more likely to have had breaks or delays in chemotherapy, 6.1 times more likely to have had their dose of chemotherapy reduced, and nearly twice as likely to have received feeding tubes or TPN (for reasons other than prophylaxis) or indwelling i.v. lines, although all the corresponding 95% CIs overlapped 1.0.

DISCUSSION

Virtually all patients who receive radiation therapy for HNC develop some degree of OM. In randomized controlled trials examining the use of conventional versus altered fractionated radiation therapy or combined radiation and chemotherapy, severe OM has been reported in up to 60% of patients receiving standard radiation therapy, and as many as 100% of those receiving either hyperfractionation or accelerated hyperfractionation regimens.2, 11-24 In the current study, the overall incidence of OM was 83%, which is consistent with the estimate of 80% reported by Trotti et al. in a recent meta-analysis of randomized trials of radiation therapy with or without chemotherapy in patients with HNC.2

The degree and duration of mucositis among irradiated patients has been reported to be directly related to the total dose and fraction, the volume of tissue treated, and the overall treatment time.5 Factors that were found to be significantly associated with an increased risk of OM in the current study included nasopharyngeal or oropharyngeal tumor location, a cumulative radiation dose ≥ 5000 cGy, and concomitant chemotherapy. However, we did not find that the use of hyperfractionated or accelerated radiation therapy conferred an increased risk of OM, contrary to the Radiation Therapy Oncology Group (RTOG) Phase III trial (RTOG 9003), in which all three altered fractionation regimens (i.e., hyperfractionation, accelerated fractionation with split, and accelerated fractionation with concomitant boost) were reported to result in more acute toxicity (-primarily mucositis) than standard fractionation.25 In RTOG 9003, the incidence of Radiation Morbidity Scoring Criteria Grade 3 or 4 OM occurring within 90 days of treatment was found to be 25% for standard fractionation, 41.8% for hyperfractionation, 40.9% for accelerated fractionation with split, and 45.9% for accelerated fractionation with boost. In the current study, 76% of those patients receiving hyperfractionation and 58% of those receiving hyperfractionation plus standard fractionation experienced moderate or severe OM compared with 65% for those receiving standard fractionation. However, it should be noted that relatively few patients in the current study received altered fractionation regimens (hyperfractionation in 26 patients, hyperfractionation plus standard fractionation in 43 patients, and accelerated fractionation in 7 patients). Although we acknowledge the possibility that the incidence and severity of OM could have been consistently underreported for patients who received hyperfractionated or accelerated regimens, we suspect that our findings would have been similar to those reported in RTOG 9003 had greater numbers of patients in the current study received such treatment.

The incidence of OM has been reported to vary not only by treatment but also in relation to various patient characteristics, including advanced age, gender, smoking and alcohol consumption, altered oral intake, preexisting periodontal disease, low body mass index, poor functional status, low leukocyte count, advanced disease and stage, a prior history of severe mucositis, and various comorbid conditions. However, the data supporting each of these potential risk factors are uneven. Although we did not comprehensively evaluate all of the above-mentioned measures, advanced age was found to confer a lower risk of OM (in contrast to prior research), although this finding may reflect a lower intensity of treatment that could not be adequately controlled for. A prior history of smoking and alcohol consumption were not found to be associated with a higher incidence of OM on multivariate analyses. We also found that indemnity insurance was a predictor of OM, although the significance of this finding is unclear.

Data from small studies3, 6 and meta-analyses2 suggest that mucosal injuries may result in breaks in treatment, the placement of feeding tubes, and hospitalization. The current study findings support these hypotheses. In our study, patients with OM were fourfold more likely to have had unplanned breaks in radiation therapy and more than three times as likely to have been hospitalized. Although we did not address the clinical events leading to these outcomes, patients with severe mouth and throat pain may require injectable narcotics, develop streptococcal or Candida infections as a result of the breakdown of oral mucosa, and have poor nutritional status and weight loss as a result of their inability to eat or drink.4, 7, 27 These conditions may persist for months and add significantly to the burden and cost of treatment.

A few limitations of the current study should be noted. First, because the current study was retrospective and designed specifically to include patients from a broad range of practice settings (academic as well as community-based), we did not attempt to limit OM assessment to a specific scale. Although various standardized instruments are available for grading OM, the occurrence of OM and its severity were ascertained based on provider assessment and any notation indicating whether the condition was mild, moderate, or severe. The extent to which investigators employed similar definitions of mild, moderate, and severe OM is unknown. Second, because medical records were not randomly selected for abstraction, a potential for selection bias exists; prior knowledge of information pertinent to the study may have prompted investigators to abstract data for some patients and not others. Third, because the current study was retrospective and involved the concurrent assessment of OM severity, potential risk factors for mucosal toxicity, and adverse outcomes, we cannot rule out the possibility that this imparted a bias to our findings, especially with respect to the correlation between OM and the outcomes of interest. Investigators, for example, may have been more likely to report moderate or severe OM in patients who had been hospitalized or had breaks/delays in their treatment. Finally, missing data limited our ability to consider additional explanatory factors in our analyses of OM risk, such as the timing of radiation therapy and concomitant chemotherapy, the volume of the oral cavity within the radiation field, the total dose to each of the accessible sites, and the individual chemotherapeutic agents and dosages that patients received.

Despite these limitations, we believe the current study confirms that OM is a frequent complication among HNC patients, and that it is often associated with adverse clinical outcomes such as breaks/delays in treatment and hospitalization. Although other authors have described the association between OM severity and adverse outcomes, and have emphasized the need for the inclusion of such measures in future clinical trials,2 to our knowledge, the current study is the first to report the magnitude of this association in a broad range of care settings and outside the context of a clinical trial. Our findings that patients with OM are fourfold more likely to have unplanned breaks in radiation therapy and more than three times as likely to be hospitalized are extremely relevant when the setting of care is considered. Other findings (notably, that patients with OM also are nominally more likely to have breaks or delays in chemotherapy, chemotherapy dose reductions, and to receive feeding tubes or TPN [for reasons other than prophylaxis] or indwelling i.v. lines) suggest that a prospective, longitudinal cohort study in this patient population (with a standardized assessment of risk factors, mucosal toxicity, and outcomes) would be most valuable.