Impact of diabetes mellitus in patients undergoing radiation therapy for head and neck cancer Authors: Silpa Johnson, Sandeep Muzumder, Hadrian Noel Alexander F, Nirmala Srikantia, Avinash H Udayashankara, Swathi R. Krishna, Deepu C Tom, John M. Raj Published: 2026 ABSTRACT BackgroundThe study aims to find the impact of diabetes mellitus (DM) in patients undergoing radiation therapy (RT) for head and neck cancer (HNC).Materials and methodsThe retrospective cohort study analysed 346 HNC patients who received definitive or adjuvant RT with or without chemotherapy from January 2013 to December 2023. The patients were categorized into DM (n = 84) and non-DM (n = 262) group. This study analysed baseline characteristics, compliance, toxicities, and overall survival (OS).ResultsA total of 12 (14.3%) patients in the DM arm and 24 (9.2%) in the non-DM arm did not complete the scheduled RT. Treatment breaks (> 2 days) occurred in 49 patients (58.3%) in the DM arm vs. 87 (33.2%) in the non-DM arm. Grade ≥ 3 acute toxicities were significantly frequent in the DM arm, including mucositis (61.9% vs. 42.7%), dysphagia (44.0% vs. 26.0%), pain (53.6% vs. 40.1%), febrile neutropenia (34.5% vs. 24.4%), and infections (66.7% vs. 54.2%). Supportive interventions were also significantly higher in the DM arm: antibiotics (60.7% vs. 45.4%), antifungals (51.2% vs. 39.7%), feeding procedures (33.3% vs. 16.8%), hospitalizations (58.3% vs. 40.1%), and Intensive care unit (ICU) admissions (4.8% vs. 2.3%). After 17.15 months’ follow-up, mortality was 47.6% in DM vs. 35.1% in non-DM (p = 0.040), with median OS 35.3 months in DM and not reached in non-DM. Restricted mean survival was significantly higher in the non-DM group (42.3 vs. 34.3 months; p = 0.0059).ConclusionsThis study showed that HNC patients with DM undergoing RT are more susceptible to acute and late toxicities, poor compliance to treatment and significantly poor OS compared to non-DM group. Introduction Head and neck cancers (HNCs) represent a major global health challenge, with incidence rates steadily rising over the recent decades [1]. Annually, approximately 900,000 new cases and over 400,000 deaths are reported worldwide [2]. Radiation therapy (RT) remains a cornerstone in the management of HNC, used either as a definitive or adjuvant therapy after surgery, with or without chemotherapy [3, 4]. RT improves local control (LC) and overall survival (OS), and enables organ preservation in some cases, which helps to minimize the impact on the patient’s quality of life [5]. However, RT may result in several undesired reactions like mucositis, pain, dysphagia, aspiration and infection [6]. Factors that increase the side effects of RT include altered fractionation, larger fields, concurrent chemotherapy (CRT), advanced disease stages, poor general condition of the patient, poor nutritional status, and comorbidities like diabetes mellitus (DM) [7, 8]. In cancer patients, 8–18% have DM as a co-morbid condition, with prevalence varying by cancer site [9]. The prevalence of DM patients with HNC has been reported to range from 12.8% to 18.5% [10–12]. Also, the incidence of HNC in patients with newly diagnosed DM is 1.47 times higher compared to those without diabetes [13]. As a frequent co-morbidity of cancer patients, DM is reported to compromise the oncology outcome [14]. Several studies showed that DM significantly worsens complications in HNC patients undergoing CRT, including higher rates of infection, haemato-toxicity, weight loss, and mortality [8, 10, 15]. Additionally, DM exacerbates acute toxicities from RT and CRT, particularly when blood glucose and hypertension (HTN) are poorly controlled [16]. Understanding how DM impacts the efficacy and side effects of RT in HNC patients is essential for optimizing treatment strategies and improving patient care. This study aimed to evaluate the impact of DM on treatment compliance, toxicity and survival outcomes in HNC patients undergoing RT or CRT at our institution. Materials and methods Study design and patient population After Institutional ethical clearance (176/2024), all HNC patients undergoing RT at the Department of Radiation Oncology, St. John’s Medical College, Bangalore, from January 2013 to December 2023 were analysed for this retrospective study. The inclusion criterion required histologically proven diagnosis of HNC patients undergoing definitive or adjuvant RT with or without chemotherapy. Patients who received palliative RT or those who underwent re-irradiation were excluded. A total of 346 eligible patients were identified. The study by Kuo et al. [10] showed that there is a 23% difference in acute radiation induced toxicity (23 vs. 46%) between non-diabetic and diabetic patients receiving RT. To observe this difference with 80% power and 5% level of significance the sample size required is 72 per group. Staging followed the criteria of the American Joint Committee on Cancer (AJCC) 8th edition (2018). The collected parameters included baseline patient demographics, disease stage, tumour histology, planned treatment, treatment received, early and late side effects and interventions done during treatment. Based on their documented medical history, patients were categorized into diabetic (DM) and non-diabetic (non-DM) groups. Diabetes mellitus was defined as a previously documented diagnosis of type 1 or type 2 diabetes in the medical records or current use of antidiabetic medications [17]. The performance status of the patients was assessed according to the Eastern Cooperative Oncology Group (ECOG) performance scale. Toxicity grading was performed using Common Terminology Criteria for Adverse Events (CTCAE) 5.0. All patients were evaluated twice weekly during the treatment period with weekly complete blood count and after treatment weekly till acute toxicity subsided, and then at every monthly follow up. All acute and late toxicities, along with any treatments administered, were documented. Acute toxicities were observed within 90 days from the start of RT. Patients who discontinued treatment or had treatment breaks due to acute toxicity were analyzed separately and were not grouped with other default cases. Patients who discontinued treatment on their own without a due reason were defined as defaulters. Patients whose treatment was interrupted due to external factors, such as machine failure, accidents, or unrelated medical conditions, were defined as Logistic, and patients who were unable to complete treatment due to financial constraints were defined as Others. Treatment response was assessed using cross-sectional imaging with computed tomography (CT) or magnetic resonance imaging (MRI) at 3 months after completion of treatment. Clinical evaluations were performed at each follow-up visit, and any abnormal findings prompted earlier imaging for response assessment. Follow-up information from the last clinic visit was obtained from hospital records. Patients who missed follow-ups were contacted to inquire about their well-being, and outcomes were recorded as either alive or deceased. Study outcome The primary objective of the study is to compare the incidence and severity of radiation-related toxicity in DM vs. non-DM arms. Secondary objectives included comparison of treatment compliance, progression-free survival (PFS) and OS. PFS was calculated from time of diagnosis to an event like recurrence (loco-regional or distal), progression, second primary or death due any cause. Disease progression was defined as the appearance of new lesions or a progressive increase in tumor size during treatment or immediately after completion of therapy without prior complete response. Recurrence was defined as the reappearance of lesions following an initial complete response, either locally, regionally, or at distant sites. OS was defined from the date of diagnosis to date of death. The restricted mean survival time was a time-dependent measure that reflects the average survival duration over a defined period with adequate follow-up, hence the term. Statistical analysis All categorical data were presented using frequency and percentages, and all continuous data using the mean and standard deviation or median and interquartile range (IQR) based on the distribution. The differences in toxicities between the two arms were compared using the chi-square test. PFS and OS were examined using Kaplan–Meier survival plots and compared using log-rank tests. The restricted mean survival time was used as the survival metric when the number of events was insufficient to calculate median survival reliably. The Cox proportional hazard model was used to examine the difference between groups for OS after adjusting for confounding variables. A p-value < 0.05 was considered statistically significant. Results Baseline characteristics A total of 346 patients with HNC who received RT between January 2013 and December 2023 were included in this study. Of these, 84 and 262 patients were in DM and non-DM group, respectively. Among the 84 diabetic patients, 58 (69.0%) were on oral hypoglycemic agents, 12 (14.3%) were on insulin therapy, and 14 (16.7%) were receiving a combination of oral hypoglycemics and insulin. In the DM group, the median pre-RT glycated haemoglobin (HbA1c) was 8.1 (IQR: 6.8–9.5), and post-RT random blood sugar (RBS) in a sub-sample was 123 mg/dL (IQR: 100–148). In the non-DM group, the median pre-RT RBS was 97 mg/dL (IQR: 89–111). Their baseline characteristics are presented in Table 1. Most patients in both groups were male with a good performance status. About 50% of the DM group had HTN, compared to 18.7% in the non-DM group. Tobacco use was reported in 53.44% of the non-DM group and 58.3% of the DM group, while alcohol use was observed in 34.5% of the DM group and 32.4% of the non-DM group. The cancers of oral cavity were most common site of primary followed by carcinoma of larynx in both the groups. Most patients in both arms of the study had locally advanced disease and, therefore all patients received a multi-modal therapy. Treatment compliance The median RT dose delivered was 6600cGy in both arms. The planned RT dose was completed by 90.8% of patients in the non-DM group compared to 85.7% in the DM group (p = 0.18). Among those who did not complete their planned therapy, toxicity was a more frequent reason in the DM group (66.7%) than in the non-DM group (41.7%). Forty-nine (58.3%) patients in the DM group experienced breaks (> 2 days) in RT, compared to 87 (33.2%) in non-DM (p < 0.001). Toxicity accounted for treatment break in 87.8% in the DM and 59.8% in Non-DM group (p = 0.003). Most patients were treated by the Intensity Modulated Radiotherapy technique in either arm. The proportion of patients receiving chemotherapy was similar between the two groups. Majority of patients in both arms received weekly concurrent cisplatin chemotherapy. Toxicity was the primary reason for not receiving planned chemotherapy in both groups (88%). The median duration of radiotherapy was 41 days in the DM group and 42 days in the non-DM group. Table 2 provides additional details regarding treatment compliance. Acute and late toxicities The incidence of acute toxicities and interventions required for its management are presented in Table 3. The patients in the DM group experienced higher rates of Grade ≥ 3 acute toxicities compared to the non-DM arm: pain (53.6% vs. 40.1%), dysphagia (44% vs. 26%), mucositis (61.9% vs. 42.7%), dermatitis (15.5% vs. 6.1%) and other infections (66.7% vs. 54.20%), all p < 0.05. The DM group had higher rates of antibiotic (60.7% vs. 45.4%, p = 0.015) and antifungal use (51.2% vs. 39.7%, p = 0.064). They also required feeding procedures like Naso-gastric tube insertion or feeding jejunostomy more often (33.3% vs. 16.8%, p = 0.001). Feeding tube dependence was evaluated at 1, 3, 6 months post-RT. In the diabetic group, 26 patients required tube feeding up to 1 month post-RT and 2 patients up to 3 months. In the non-diabetic group, 33 patients required tube feeding up to 1 month, 10 up to 3 months, and 1 patient up to 6 months post-RT. No patients required tube feeding beyond 6 months in either group. The DM group had a significant weight loss, higher hospitalization rate (58.3% vs. 40.1%) and also stayed at hospital longer than the non-DM group. The incidence of late grade ≥ 3 toxicities was significantly more common in the DM group compared to the non-DM group: aspiration occurred in 2.4% vs. 0% (p = 0.043), dysphagia in 4.8% vs. 0.4% (p = 0.007), and xerostomia in 2.4% vs. 0% (p < 0.001). Acute aspiration was identified clinically by symptoms such as coughing or choking during swallowing and confirmed with chest X-ray when indicated. Survival outcomes The median follow-up period for the entire cohort was 17.15 months. The events for PFS occurred in 146 patients: 37 (44.0%) in the DM group and 109 (41.6%) in the non-DM group (p = 0.69). The median PFS was higher in the non-DM group but not statistically significant (58.9 vs. 26.6 months; p = 0.24) (Fig. 1). There were 132 deaths (38.2%) in the cohort: 40 (47.6%) in the DM group and 92 (35.1%) in the non-DM group (p = 0.040). The median OS in the DM group was 35.3 months, whereas the median OS was not reached for patients in the non-DM group (Fig. 2). The restricted mean survival was significantly higher in the non-DM arm (42.3 months) compared to the DM arm (34.3 months) (p = 0.0059). On Cox univariate analysis, primary tumor site (oral cavity and oropharynx), nodal stage, and radiotherapy completion were significantly associated with OS (Tab. 4). In multivariate analysis, only primary site of disease (oral cavity and oropharynx) and RT completion remained significant predictors (Tab. 4). Discussion The study found that HNC patients with DM undergoing RT had lesser treatment compliance compared to patients without DM. They also have increased incidence of grade 3 or more acute mucositis, pain, dysphagia, dermatitis, febrile neutropenia, aspiration, and infections. DM patients needed more antibiotics, antifungals, growth factors, feeding procedures, hospitalizations, ICU admissions, and had higher death rates. They also experienced more late side effects like dysphagia, aspiration, and xerostomia. Both PFS and OS were worse in the DM group, with the difference reaching statistical significance for OS. The study found that HNC patients with DM undergoing RT had lesser treatment compliance compared to non-DM patients, likely due to the higher incidence of grade 3 or more acute mucositis, pain, dysphagia, dermatitis, febrile neutropenia, aspiration, and infections. These toxicities often led to treatment interruptions, hospitalizations, and supportive interventions, thereby compromising treatment completion. Additionally, DM patients experienced higher rates of late toxicities, including dysphagia, aspiration, and xerostomia. Consequently, both PFS and OS were inferior in the DM group, with the difference reaching statistical significance for OS. Completing the RT course without interruption and maintaining overall treatment time is important for LC and OS in HNC [18]. In our study we found that 58.3% of DM patients experienced RT breaks compared to 33.2% in non-DM patients (p < 0.001), with a significant association between DM and higher treatment interruptions. Toxicity accounted for most of these breaks in the DM group. DM patients are more prone to developing acute radiation-related side effects, which can disrupt treatment schedules [5]. This is consistent with existing literature, which shows that DM often leads to increased radiation toxicity, affecting treatment adherence [5, 8]. Cannon et al. [19] noted that treatment prolongation, especially in HNC, is associated with worse outcomes, including higher local failure rates [19]. However, despite these interruptions, the rate of RT completion (85.71% for DM patients) was relatively high, comparable to non-DM patients (90.84%). This aligns with studies indicating that although DM can complicate treatment effects, with proper management, diabetic patients can still complete prescribed RT course [8]. Diabetes mellitus has been associated with increased susceptibility to infections [20]. In this study DM patients experienced more severe early and late toxicities compared to non-DM patients. Several studies also reported increased severe toxicity in DM patients undergoing CRT, including infections and mucositis [8, 10, 16]. Erdiş et al. [16] also highlighted the increased risk of severe side effects when DM coexists with HTN, which was also a trend observed in our study where 50% of diabetic patients had both conditions. Our study also showed similar trends. DM patients had higher rates of severe mucositis (61.9% vs. 42.74% in non-DM) and infections (66.67% vs. 54.20%) in the present study. The rate of hospitalizations for toxicity was higher in the DM group compared to the non-DM group, which is consistent with previous studies. Our research showed that DM patients required more hospitalizations, ICU admissions, and enteral tube feeding than non-diabetic patients. This finding aligns with previous studies highlighting the increased care needs of diabetic patients undergoing RT [8, 10]. Diabetic patients undergoing treatment for HNC face higher treatment-related mortality rates than non-diabetic patients. In our study, 7.14% of diabetic patients died from treatment complications, compared to only 1.91% in non-diabetic patients. This finding is consistent with Foreman et al. [12], who noted that diabetes significantly influences cancer mortality due to increased complications. In our study, DM did not show a significant association with PFS. This finding is consistent with existing literature by Kuo et al. [10], which showed no significant association between DM and disease-free survival. However, DM was associated with worse OS, supporting previous research by Wu et al. [15], which found poorer OS outcomes in patients with DM. DM is associated with approximately a two-fold higher risk of death from macrovascular diseases such as coronary heart disease and stroke, which may contribute to the higher non-malignant mortality (15–35% at 5 years) reported among HNC patients [21, 22]. The effects of DM on survival outcomes in HNC remain inconsistent across studies. Some research showed that DM had negative impact on the survival of patients with HNC undergoing multimodality treatment, including surgery, chemotherapy, and radiotherapy [15, 23]. In contrast, Foreman et al. [12] showed that DM alone did not adversely affect cancer survival outcomes of patients with HNC [8]. Similarly, Spratt et al. [24] found no difference in loco-regional control between patients with and without DM in oropharyngeal cancer [8]. A meta-analysis study comparing OS in cancer patients with and without pre-existing DM showed that DM was associated with an increased hazard ratio of 1.4 for all causes of mortality, which also aligns with our findings [25]. We acknowledge that our study has certain limitations. Firstly, this is a retrospective study. Second, the data comes from single institutional experience and may not represent common practice. Third, blood sugar values were not monitored when dexamethasone was administered as an antiemetic during treatment time which might have affected glucose control. In addition, diabetic patients were not stratified based on the degree of glycemic control (e.g., HbA1c levels), which could have influenced treatment tolerance and outcomes. Lastly, we did not analyse the exact causes of death, which limited understanding of the impact of DM on mortality. Our study has several strengths. First, it’s based on real-world data showing results from every day clinical practice. Despite a relatively short median follow-up of 17.15 months, our study provides valuable insights into early outcomes and treatment-related toxicities in this population. Most importantly, it is the first study from India to look at how diabetes affects patients with HNC undergoing RT. Diabetes management during cancer treatment requires a multidisciplinary team, including oncologists and diabetes specialists, to provide comprehensive care [26]. Screening for baseline blood glucose or HbA1c levels, monitoring metabolic control, and assessing diabetic complications are essential [8, 26]. Patients at higher risk should undergo regular monitoring, including fasting plasma glucose checks and HbA1c measurements at scheduled intervals [26]. Advanced tools such as continuous glucose monitoring can aid in preventing severe hyperglycemia and hypoglycemia. Individualized treatment plans, close glucose monitoring, lifestyle modifications, and effective communication among healthcare providers are critical for optimal management and improved treatment outcomes [26]. Conclusions In our study, HNC patients with DM undergoing RT were more susceptible to acute and late toxicities, had poorer treatment compliance, and demonstrated significantly reduced OS compared to non-DM patients. Further prospective studies are warranted to better understand the impact of diabetes on oncologic outcomes.