Abstract
Background/Aim: Smoking and alcohol abuse may impair outcomes of chemoradiation for squamous cell head and neck cancer (SCCHN). Potential associations with toxicity, loco-regional control (LRC), and overall survival (OS) were investigated. Patients and Methods: Ninety-six patients were retrospectively analyzed for impacts of pre-radiotherapy (pre-RT) smoking history, smoking during radiotherapy, and pre-RT alcohol abuse on toxicity, LRC, and OS. Results: A trend was found for associations between pre-RT smoking history and grade ≥2 dermatitis. Smoking during radiotherapy was significantly associated with grade ≥3 mucositis and showed trends regarding grade ≥2 mucositis and dermatitis. On univariate analyses, smoking during radiotherapy was negatively associated with LRC and OS, pre-RT alcohol abuse with OS, and >40 pack years with LRC and OS. In multivariate analyses, smoking during radiotherapy remained significant for decreased OS, and pack years showed a trend. Conclusion: Smoking during radiotherapy was an independent predictor of OS and associated with increased toxicity. Thus, it is important to stop smoking prior to the start of radiotherapy.
Many patients treated with chemoradiation for squamous cell head and neck cancer (SCCHN) have a history of smoking and alcohol use (1–3). A pre-radiotherapy (pre-RT) history of smoking, namely the number of pack years, was reported by several authors to be associated with worse outcomes including decreased local control (LC) and overall survival (OS) (4–8). Moreover, a negative impact of a greater number of pack years on treatment outcomes were also reported for patients irradiated for lung cancer or carcinoma of the urinary bladder (9–11). However, in a population-based study of 316 patients with larynx cancer, no significant associations were found between pre-RT history of smoking and outcomes in terms of disease-free survival (DFS) and OS (12). The results of available studies in patients irradiated for SCCHN regarding the impact of smoking during a course of radiotherapy and treatment outcomes were more heterogeneous. A few studies showed significant associations between smoking during the radiotherapy course and treatment outcomes (5, 13, 14), whereas other studies did not find such correlations (8, 12, 15). The results of previous studies in patients irradiated for SCCHN were heterogeneous also with respect to the impact of smoking on treatment-related toxicity. Several studies found significant associations between smoking history or smoking during radiotherapy and increased acute toxicities (16–19), whereas other studies and a systematic review did not show such associations (8, 13, 14, 20).
Thus, additional studies are required to better define the impact of smoking prior to and during a course of radiotherapy on toxicity and other treatment outcomes in patients with SCCHN. The present study was performed to help answer this important question. In addition, the potential role of a history of alcohol abuse was investigated. The study focused particularly on patients treated with concurrent cisplatin-based chemoradiation, who are at a higher risk of experiencing treatment-related toxicities than patients receiving radiotherapy alone (21, 22).
Patients and Methods
This retrospective study (Ethics Committee of the University of Lübeck, 21-034) included data of 96 patients with SCCHN, who were treated with concurrent chemoradiation between 2010 and 2020. Sixty-six of these patients were included in a previous study (23). Median total dose of conventionally fractionated external beam radiotherapy (EBRT), mostly performed as volumetric modulated arc therapy, was 66 Gy (range=60–70 Gy). Nine patients received a brachytherapy boost in addition to EBRT, ranging between 7.5 and 16 Gy in 3 to 5 fractions. Concurrent chemotherapy was planned to include two courses of cisplatin alone, each course consisting of 20 mg/m2 on days 1–5 or 25 mg/m2 on days 1–4. Thus, the planned cumulative cisplatin dose was 200 mg/m2.
The major goal of this study was to evaluate potential associations between pre-RT history of smoking (no vs. yes), smoking during the radiotherapy course (no vs. yes), and pre-RT history of alcohol abuse (no vs. yes) and outcomes in terms of toxicity, loco-regional control (LRC), and OS. Alcohol consumption during the radiotherapy course was not properly documented in many patient files and, therefore, could not be studied. Investigated grade ≥2 and ≥3 toxicities included oral mucositis, radiation dermatitis, xerostomia, subcutaneous fibrosis, and cervical lymph edema. Oral mucositis was graded according to the Radiation Therapy Oncology Group (RTOG) criteria, other toxicities according to Common Terminology Criteria for Adverse Events version 5.0 (CTCAE v5.0) (24, 25).
Regarding LRC and OS, 11 patient-, tumor- or treatment-related characteristics were evaluated in addition to smoking and alcohol consumption. These characteristics included age (≤60 vs. >60 years), sex (female vs. male), Karnofsky performance score (KPS ≤80 vs. 90–100), cancer site (oropharynx vs. hypopharynx vs. larynx vs. oral cavity/floor of mouth vs. more than one site), primary tumor stage (T1-3 vs. T4), nodal stage (N0-2a vs. N2b-3), histologic grading (G1-2 vs. G3), human papilloma virus (HPV) status (negative vs. positive), upfront surgical resection (no vs. yes), radiotherapy breaks >1 week (no vs. yes), and cumulative cisplatin dose (≤180 mg/m2 vs. >180 mg/m2). The distribution of these characteristics is shown in Table I. Moreover, in 48 patients with pre-RT history of smoking, the number of pack years was available and analyzed.
Statistical analyses regarding toxicities were performed with the Fisher’s exact test. p-Values <0.05 were considered indicating significance, p-values <0.20 a trend. Univariate analyses of LRC and OS rates were performed with the Kaplan-Meier method and the log-rank test. LRC and OS were calculated from the day of the last radiation fraction. Again, p-values <0.05 were considered indicating significance, p-values <0.20 a trend. Characteristics significantly associated with LRC or OS were subsequently analyzed for independence using a Cox proportional hazards model (multivariate analyses). Additional multivariate analyses were performed limited to characteristics that were available for at least 90 patients. The same criteria regarding significance and trends were applied as for the univariate analyses.
Results
The comparison of patients with a pre-RT history of smoking and those patients without such a history revealed no significant differences with respect to the rates of the investigated toxicities (Table II). A trend was found for an association between pre-RT history of smoking and grade ≥2 radiation dermatitis (86% vs. 72%, p=0.16). Moreover, smoking during the radiotherapy course was significantly associated with grade ≥3 oral mucositis (37% vs. 14%, p=0.012) and showed trends for associations with grade ≥2 oral mucositis (80% vs. 64%, p=0.11) and grade ≥2 radiation dermatitis (91% vs. 79%, p=0.15) (Table III). No significant associations were found between pre-RT history of alcohol abuse and treatment-related toxicity (Table IV).
On univariate analyses of LRC, smoking during the radiotherapy course was significantly associated with worse outcomes (p=0.020) (Table V). No such associations were found for pre-RT history of smoking (p=0.90) and pre-RT history of alcohol abuse (p=0.26). In contrast, improved LRC was significantly associated with favorable cancer site (oropharynx or larynx, p=0.013), T1-3 stage (p<0.001), N0-2a stage (p=0.046), HPV-positive tumors (p=0.026), and upfront resection (p<0.001). In addition, trends were found for KPS 90–100 (p=0.17) and cumulative cisplatin doses >180 mg/m2 (p=0.19). In the subsequent multivariate analysis, upfront surgery [hazard ratio (HR)=0.21, 95% confidence interval (CI)=0.04–1.13, p=0.069] and T-stage (HR=3.37, 95%CI=0.62–18.23, p=0.16) showed trends for associations with LRC. N-stage (HR=1.49, 95%=CI 0.69–3.23, p=0.31), smoking during radiotherapy (HR=1.94, 95%CI=0.65–5.81, p=0.23), cancer site (HR=1.03, 95%CI=0.80–1.32, p=0.82), and HPV-status (HR=0.33, 95%CI=0.03–3.12, p=0.33) were not significant in the multivariate analyses of LRC.
In an additional univariate analysis in those 48 patients with pre-RT history of smoking and available number of pack years, >40 pack years were significantly associated with worse LRC (p=0.049). In the corresponding additional multivariate analysis, pack years were not significant (HR=1.33, 95%CI=0.13–13.18, p=0.81).
On univariate analyses of OS, smoking during the radiotherapy course (p<0.001) and pre-RT history of alcohol abuse (p=0.025) were significantly associated with worse outcomes (Table VI). No association was observed for pre-RT history of smoking (p=0.31). Moreover, OS was positively associated with KPS 90–100 (p=0.020), HPV-positivity (p=0.009), and cumulative cisplatin doses >180 mg/m2 (p=0.041). In the multivariate analyses, smoking during the RT course was significant (HR=4.80, 95%CI=1.38–16.67, p=0.013), and a trend was found for HPV-status (HR=0.30, 95%CI=0.05–1.68, p=0.17). Cumulative cisplatin dose (HR=0.52, 95%CI=0.19–1.45, p=0.21). KPS (HR=0.64, 95%CI=0.21–1.97, p=0.44), and pre-RT history of alcohol abuse (HR=1.37, 95%CI=0.44–4.25, p=0.59) did not achieve significance in the multivariate analysis. In the additional analysis of the 48 patients with pre-RT history of smoking and available number of pack years, >40 pack years were significantly associated with decreased OS (p=0.027) on univariate analysis and showed a trend in the multivariate analysis (HR=8.72, 95%CI=0.76–100.55, p=0.083).
Discussion
Despite the contribution of many studies performed during recent years, the outcomes of patients receiving radiotherapy or chemoradiation for SCCHN need to be improved (26–31). The prognosis may be impaired by several lifestyle factors including smoking and alcohol consumption (1–3). Available studies had produced conflicting results regarding the impact of cigarette smoking and alcohol on treatment-related toxicities and outcomes in terms of LC, LRC, and OS (4–8, 12–20). Several studies found that a greater number of pack years prior to the start of radiotherapy was associated with worse outcomes (4–8). In the study of Oliva et al., the number of pack years was negatively associated with OS (HR=1.14, p=0.01) but not with cancer-specific survival (4). In two other studies, a greater number of pack years was associated with significantly decreased OS and recurrence-free survival (RFS) or with significantly decreased OS and progression-free survival (PFS), respectively (6, 7). However, in the population-based study of Sjögren et al. that included 316 consecutive patients irradiated for T1-stage glottic cancer, smoking history showed no significant impact on LC, DFS, and OS (12). Five-year LC rates were 89% in non-smokers, 94% in ex-smokers, and 84% in smokers, respectively (p=0.27). Data were even more conflicting regarding associations between smoking during the radiotherapy course and subsequent outcomes. In the study of Vawda et al., patients smoking during the radiotherapy course had significantly worse 5-year OS rates than former smokers and never smokers (62% vs. 87% and 76%, respectively; p<0.001) (5). Five-year RFS rates were also significantly worse in that study (72% vs. 86% and 83%, respectively; p=0.03). In another study and in a systematic review, smoking during the radiotherapy course resulted in significantly worse OS and response and in worse LRC and OS, respectively (13, 14). In contrast to these studies, three other studies did not find significant associations between smoking during the radiotherapy course and treatment outcomes including LC and OS (8, 12, 15). When considering these conflicting results, it becomes obvious that additional studies are needed to clarify the potential impact of smoking prior to and during radiotherapy for SCCHN.
This need is further supported by examining the available data regarding the impact of smoking on treatment-related toxicities. In several studies, smoking prior to or during a course of radiotherapy was associated with increased rates of acute toxicities and complications including oral mucositis, skin reactions, need for enteral feeding, and hospitalization (16–19). Moreover, one study and a systematic review reported higher rates of late toxicities (13, 19). However, the systematic review and three previous studies did not find significant associations between smoking and acute toxicities (8, 13, 14, 20).
The current study was conducted to contribute to further clarification of the prognostic role of smoking on toxicity, LRC and OS. In order to reduce the risk of a treatment-related selection bias, only patients who were scheduled for concurrent chemoradiation with two courses of 20 mg/m2/d1-5 or 25 mg/m2/d1-4 of cisplatin were included. According to the results of this study, a greater number of pack years and smoking during radiotherapy were negatively associated with LRC and OS. Moreover, smoking during radiotherapy was significantly associated with grade ≥3 mucositis and trends were found for associations between pre-RT smoking history and grade ≥2 dermatitis and for smoking during radiotherapy and grade ≥2 mucositis and grade ≥2 dermatitis. These results agree with the findings of some previous studies and support the idea that it is important to stop smoking prior to the start of radiotherapy to improve LRC and OS and reduce acute toxicities (4–8, 13, 14, 16–19). Reduction of acute toxicities is particularly important, since adverse effects may lead to interruption of the radiation treatment. Such interruptions were shown to impair patient prognoses in terms of decreased LRC and OS (32). Furthermore, continuation of smoking during a course of radiotherapy was shown to have a negative impact on the patients’ quality of life (33, 34).
The fact that smoking during radiotherapy has a negative impact on treatment outcomes can be explained to a certain extent by the fact that smoking increases the levels of carboxy-hemoglobin, which is known to impair the delivery of oxygen to the tumor cells (35–37). Hemoglobin binds to carbon monoxide significantly stronger (approximately by factor 200) than to oxygen. Thus, hemoglobin cannot take sufficient oxygen to the tumor cells. Oxygenation of tumor cells is very important for radiotherapy-induced tumor-cell killing, which is mediated through cytotoxic free oxygen-radicals (38). Associations between biomarkers of poor oxygenation and poor prognoses were previously described in patients with head and neck cancer (39). These explanations are supported by the results of the present study. Moreover, an enhanced impact of smoking and alcohol abuse on the immunological balance of monocyte subsets and CD4/CD8 T-cells in the peripheral blood of head and neck cancer patients was previously observed (40). When interpreting the results of the current study, the retrospective design bearing the risk of hidden selection biases needs to be considered. Furthermore, since all patients received concurrent chemoradiation, it is not clear whether these results are applicable to patients treated with radiotherapy alone or sequential chemoradiation.
In addition to smoking, the potential impact of a history of alcohol abuse was investigated in the current study. Pre-treatment alcohol consumption was associated with significantly worse OS on univariate analysis but not in the multivariate analysis. This may be explained by a correlation between alcohol abuse and HPV-status. Significantly more patients with HPV-negative tumors had a history of alcohol abuse than patients with HPV-positive tumors (70% vs. 16%, p<0.001; Fisher’s exact test). The studies available so far produced conflicting results regarding the prognostic role of alcohol consumption. In the study of Zanoni et al., a history of alcohol consumption was significantly associated with worse disease-specific survival in patients with cancer of the oral cavity (41). Moreover, Sawabe et al. reported a negative impact of alcohol consumption in the subgroup of head and neck cancer patients irradiated for cancer of the oropharynx, hypopharynx, or larynx (42). In contrast, in the study of Colasanto et al., OS and local recurrence-free survival were not associated with history of alcohol consumption or alcohol abuse during treatment of 190 patients receiving definitive radiotherapy for T1 or T2 larynx cancer (15). In another study of 168 patients with oropharynx cancer, the unit years of alcohol consumption were not associated with OS (p=0.55) but showed a trend for reduced PFS (p=0.057) (7). Since the results of the available studies regarding the prognostic role of alcohol consumption are conflicting, additional (prospective) studies are required for further clarification.
In conclusion, smoking during radiotherapy was an independent predictor of OS and associated with increased toxicity. Thus, it is important to stop smoking prior to the start of radiotherapy. Moreover, this factor should be considered when designing future prospective trials in patients receiving chemoradiation for SCCHN.
Footnotes
Authors’ Contributions
The study was designed and supported by all Authors. Data were collected by I.Z., and subsequently analyzed by D.R., N.Y.Y. and S.E.S. The article drafted by D.R. was reviewed and finally approved by all Authors.
Conflicts of Interest
The Authors report no conflicts of interest related to this study.
- Received December 22, 2022.
- Revision received December 31, 2022.
- Accepted January 2, 2023.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).