Inglês (pdf)
Artigo em XML
Referências do artigo
Tradução automática
Permalink
INTRODUCTION
Head and neck cancer (HNC) includes tumors of the upper respiratory/digestive tract, such as the oral cavity, oropharynx, nasal cavity, nasopharynx, hypopharynx, larynx, and salivary glands1. HNC is the seventh most common cancer worldwide and ranks fifth among the most prevalent types of cancer in men in Brazil, fourth in the Southeast region, and third in countries with a low or medium Human Development Index (HDI)2. The main causes of HNC are related to smoking, alcohol consumption, viral exposure to Human Papillomavirus (HPV), and a nutrient-poor diet, with tobacco being identified as the leading behavioral risk factor for the development of this disease5.
Malnutrition is a common condition in individuals with HNC6,7, and approximately 30.3% of this population is at nutritional risk before treatment initiation8. Several factors related to cancer patients can contribute to malnutrition, such as reduced food intake, tumor location, and symptoms arising from the disease, which trigger a hypercatabolic process and intensify the loss of skeletal muscle mass, depleting the nutritional status2,9.
Nutritional status can be monitored using anthropometric measures, such as calf circumference (CC), which is closely associated with skeletal muscle mass and directly affected by events inherent to the processes of frailty, malnutrition, and cancer cachexia10,11. Early detection of nutritional risk allows for the implementation of actions that can have significant results in the preservation or recovery of nutritional status. Nutritional screening, recommended by the European Society for Clinical Nutrition and Metabolism (ESPEN), is an essential tool to identify patients who need nutritional support early, especially in vulnerable populations, to identify patients requiring early nutritional intervention12.
Among the methods, the NRS-2002 stands out as an instrument recommended for use in hospital settings due to its quick and easy application15,16. It is frequently used in studies involving HNC patients before treatment, demonstrating its ability to detect malnutrition and nutritional risk13,17. According to the Brazilian Consensus on Oncology Nutrition (2021), the NRS-2002 can be used for both clinical and surgical patients18.
Simplified strategies can also be used to aid in the detection and diagnosis of nutritional risk, such as measuring the CC, as it is a simple, low-cost, non-invasive method with a good correlation to muscle mass19.
In light of the above, the aim of this study is to identify factors associated with nutritional risk as assessed by the NRS-2002 in HNC patients. Understanding the associated variables may facilitate a rapid approach and early detection of risk in this population. This, in turn, may contribute to better nutritional management and potentially better health outcomes.
METHODS
Study design
This is an epidemiological, analytical, cross-sectional investigation of individuals with HNC. The data represents the baseline of a longitudinal study titled “Nutritional Indicators, Mortality, and Associated Factors: A Hospital-Based Study in Individuals with HNC.” The study was conducted at a reference hospital in Brazil.
For the sample size calculation, a population of 200 patients was considered, which corresponds to the average annual number of first consultations for HNC in the outpatient clinic, with a prevalence of 50%, a sample error of 5%, and a 95% confidence interval (CI). This resulted in a minimum sample size of 132 individuals. The calculation was performed with the help of the Epidat 4.2 software.
Study location and period
Data were collected from September 2022 to January 2024. The individuals were invited to participate after the confirmation of the HNC diagnosis by the Head and Neck Surgery Medical Service, and the study was exclusive to patients from the Unified Health System.
Study population and eligibility criteria
Individuals of both sexes, aged 18 years or older, with a diagnosis of squamous cell carcinoma of the oral cavity, larynx, oropharynx, and hypopharynx, with no previous oncological treatment, were included. The cases were confirmed by histology and classified according to ICD-10-0320, based on the following topographies: base of the tongue (C01), oral cavity, other parts and unspecified parts of the tongue (C02), gingiva (C03), floor of the mouth (C04), palate (C05), other parts and unspecified parts of the mouth (C06), tonsil (C09), oropharynx (C10), pyriform sinus (C12), hypopharynx (C13), and larynx (C32). Subsequently, the ICD codes were grouped into three topographical categories: oral cavity (C01, C02, C03, C04, C05); oropharynx (C05.1, C09, C10); larynx and hypopharynx (C32, C12, C13).
Patients with recurrent squamous cell carcinoma and/or those with prior treatment, those with more than one tumor, and patients without the clinical or mental conditions to respond to the questionnaire were excluded from the study.
Data collection
A questionnaire was applied containing sociodemographic variables (sex, age, race/color, education, and income), lifestyle habits (smoking, alcohol consumption, and physical activity), clinical data (topography and staging), and changes in the consistency of consumed foods. Anthropometric measurements (weight, height, and CC) were also taken. It is important to note that sociodemographic variables, lifestyle habits, and changes in food consistency were collected through self-report.
Nutritional impact symptoms (SIN) screening was conducted using the “Head and Neck Symptoms Checklist”21, which consists of 17 SIN evaluated on a five-point Likert scale, ranging from “1, none” to “5, very much.” For analysis, the scores of the 17 symptoms were summed to obtain a total score, ranging from 17 (no symptoms) to 85 (maximum score). In addition, nutritional risk screening was performed using the NRS-200215, an instrument recommended by the ESPEN. The method involves two steps: the first includes questions about body mass index, weight loss, reduced food intake, and the presence of severe disease. In the second step, each criterion is quantified based on nutritional status and disease severity, with an additional point for patients aged ≥ 70 years. A total score < 3 indicates no risk, while a total score ≥ 3 indicates the presence of nutritional risk.
Data analysis
The dependent variable used was nutritional risk determined by the NRS-2002, classified as “no risk” and “at risk.” Central tendency and dispersion measures were used for continuous variables, according to the normality test. For categorical variables, absolute and relative frequencies were used. To test the association between the dependent variable and the independent variables, the Chi-square test and the Mann-Whitney test were used. The significance level for all tests was 5%.
To quantify the contribution of the independent variables to the outcome, binary logistic regression was performed, including independent variables with significance lower than 5% (p≤0.05) in the bivariate association tests. The assumption of no multicollinearity and model adjustment according to the Hosmer-Lemeshow test were considered. Odds Ratios (OR) and their respective CI were estimated.
To assess the relationship between the NRS-2002 and CC variables, Spearman’s correlation was performed, and the correlation coefficient was calculated to determine the strength and direction of the relationship between the two variables. Subsequently, the Kappa coefficient was performed to evaluate the agreement between the two variables in categorical behavior, such as NRS-2002 (“at risk” and “no risk”) and CC (“adequate” and “inadequate”).
The data were managed using the REDCap platform22 and stored in the cloud of the Federal University of Espírito Santo (UFES). Data organization and analysis were performed using the Statistical Package for the Social Sciences (SPSS®), version 22.0.
RESULTS
According to table 1, most participants were older adults (59.8%), male (77.3%), and of mixed race (54.5%). The majority lived with a partner (53.8%), had low education (71.2%), and resided in the metropolitan area of Vitória-ES (65.9%). Of the total, 75.0% had a family income of less than or equal to 2 minimum wages.
Table 1 : Sociodemographic, clinical, lifestyle, nutritional, and anthropometric data distributed according to nutritional risk in patients with head and neck cancer from a reference hospital in Brazil, 2022-2024
Chi-square test. *Mann-Whitney test. **T-test. *Data expressed as p50 ± interquartile range (IQR); **Mean and standard deviation (SD). Statistically significant difference in bold (p<0.05). N=132. 1N=127. 2N=124. SIN: Nutritional Impact Symptoms. CC: Calf circumference.
Regarding clinical data, 62.2% reported having a family history of cancer. Regarding the tumor stage at the time of diagnosis, the more advanced stages, such as III and IV, were more frequent (77.4%), and most had tumors in the oral cavity (44.7%). The majority reported past tobacco (52.3%) and alcohol (65.2%) use, with a median of 90 ± 144 packs/year for smokers and 180 ± 304.5 packs/year for ex-smokers. Physical inactivity was predominant in the sample (84.1%).
Regarding nutritional and anthropometric variables, it was observed that more than half of the sample reported a change in the consistency of foods consumed (56.8%). The average value for CC was 34 ± 4.08 cm. According to the NRS-2002, 46.2% of participants were at nutritional risk, with a median score of 2 ± 3 (table 1).
The sociodemographic, clinical, lifestyle, nutritional, and anthropometric data were analyzed according to the nutritional risk classification assessed by the NRS-2002. An association was found between nutritional risk and marital status, with a higher risk among individuals who are not living with a partner (p = 0.006). Nutritional risk was also associated with family income, particularly among those with an income equal to or less than two minimum wages (p = 0.003), changes in food consistency (p = 0.052), severity of SIN, interference in eating due to SIN, and CC (p < 0.001).
The correlation between the NRS-2002 and CC was negative and moderate, with a coefficient of -0.547. This indicates that as CC increases, nutritional risk tends to decrease. The correlation was statistically significant at the 0.001 level (two-tailed), showing a strong negative association between the NRS-2002 and CC.
The concordance between the NRS-2002 and CC methods was tested using Cohen’s Kappa Index. The Kappa value was 0.465 (p < 0.001), indicating a significant and moderate concordance between the two assessment methods, considering the NRS-2002 classification and the cutoff point for CC of ≤33 cm for women and ≤34 cm for men as indicative of inadequacy23,24. This finding suggests that both methods have moderate capacity to assess nutritional status, but they are not fully concordant.
In the multivariable analysis using binary logistic regression (table 2) with nutritional risk as the dependent variable, after model adjustment, family income and CC remained associated with the outcome. Patients with a family income of less than or equal to two minimum wages had 2.9 times higher odds (OR = 2.916; 95% CI = 1.017-8.359; p = 0.046) of developing nutritional risk compared to those earning more than two minimum wages. Regarding the anthropometric variable, it was identified that CC was also associated with nutritional risk (OR = 0.751; 95% CI = 0.646-0.873; p < 0.001), indicating an inverse relationship, where each additional centimeter in CC was associated with a 24.9% reduction in the odds of participants being at nutritional risk.
Table 2 : Multivariate analysis of nutritional risk and sociodemographic, nutritional, and anthropometric variables in patients with head and neck cancer at a reference hospital in Brazil, 2022-2024
| Variables | Unadjusted | Adjusted | ||||||
|---|---|---|---|---|---|---|---|---|
| p-value | OR | 95% CI | p-value | OR | 95% CI | |||
| Lower limit | Upper limit | Lower limit | Upper limit | |||||
| Sex | ||||||||
| Female | 1 | 1 | ||||||
| Male | 0.111 | 2.000 | 0.853 | 4.691 | 0.309 | 1.697 | 0.613 | 4.697 |
| Age | ||||||||
| Adulta | 1 | 1 | ||||||
| Older adults | 0.110 | 0.563 | 0.279 | 1.138 | 0.097 | 0.449 | 0.174 | 1.157 |
| Marital Status | ||||||||
| Living with partner | 1 | 1 | ||||||
| Not living with partner | 0.007 | 2.650 | 1.309 | 5.364 | 0.143 | 1.934 | 0.801 | 4.668 |
| Family Income | ||||||||
| ≤ 2 minimum wages | 0.005 | 0.278 | 0.114 | 0.675 | 0.046 | 2.916 | 1.017 | 8.359 |
| > 2 minimum wages | 1 | 1 | ||||||
| Dietary Consistency Change | ||||||||
| No | 1 | 1 | ||||||
| Yes | 0.027 | 2.231 | 1.098 | 4.533 | 0.428 | 1.447 | 0.580 | 3.610 |
| CC | 0.000 | 0.705 | 0.616 | 0.807 | 0.000 | 0.737 | 0.636 | 0.855 |
| Severity of SIN | 0.004 | 1.050 | 1.016 | 1.086 | 0.920 | 0.996 | 0.921 | 1.077 |
| Impact on Eating SIN | 0.001 | 1.060 | 1.025 | 1.097 | 0.227 | 1.048 | 0.971 | 1.130 |
Binary logistic regression (unadjusted and adjusted). Hosmer-Lemeshow test: 0.111; Nagelkerke R2: 0.472; Statistically significant differences in bold (p < 0.05). OR: Odds Ratio. CI: Confidence Interval. SIN: Nutritional Impact Symptoms. CC: Calf circumference. (a) Adjustment variables.
DISCUSSION
The results of the present study indicate that factors such as low income and CC are associated with nutritional risk in patients with HNC.
The sample is predominantly composed of older adults male individuals of mixed race, with low socioeconomic status, sedentary lifestyles, and a history of tobacco and alcohol use. This set of characteristics is consistent with studies addressing the epidemiology and risk factors for HNC3,25. Low socioeconomic status is a prominent feature among groups most susceptible to HNC, and it was one of the most frequent factors in the studied population, corroborating the findings of an epidemiological study conducted in an Oncological Research Center in the South of Brazil with patients with HNC26,27.
According to Barsouk et al.27, difficulty in accessing the healthcare system for disease prevention and screening is one of the factors that highlights low income as a risk characteristic for the development of HNC. Indirectly, low socioeconomic status is also associated with the main risk factors for disease development, such as smoking and alcohol consumption, with this profile showing a higher tendency to consume both27.
Contributing to the increase in nutritional risk, lower socioeconomic classes are associated with the presence of malnutrition in the oncological population, highlighting a potential link to food insecurity in individuals with low purchasing power28. Food insecurity is characterized by the lack of access to adequate and nutritious food due to financial constraints or other limiting factors29. Gajda et al.30 identified a significant relationship between food insecurity and nutritional risk in older adults individuals, demonstrating that low food security is associated with a higher nutritional risk, which may be one of the factors contributing, along with other aspects, to the findings of the present study.
Low muscle mass is a common consequence of nutritional risk and is related to longer hospitalization, lower treatment tolerance, and higher mortality risk31,32. CC is an anthropometric measurement highly correlated with direct and indirect measures of skeletal muscle mass33,34. Assuming that CC reflects skeletal muscle mass, we can explain the relationship with nutritional risk through its ability to indicate the presence of malnutrition and sarcopenia.
Ren et al.35 identified a significant relationship between CC and nutritional risk in a longitudinal study, suggesting that a low CC may be a predictive indicator of nutritional risk in patients over 80 years old. This finding is similar to what we found in our study. Although our sample is predominantly composed of older adults individuals, it also encompasses a broader age range. According to the results obtained, each centimeter increase in CC corresponds to a 24.9% reduction in the likelihood of nutritional risk. Previous studies have demonstrated that CC is a valid and reliable indicator of muscle mass19,33. Thus, a reduced CC may be an early sign of muscle mass decline.
In addition to the potential risk associated with low CC in establishing nutritional risk, we evaluated the correlation and concordance between the NRS-2002 and CC. The results showed a negative and moderate correlation (-0.547; p < 0.001) and an acceptable magnitude of concordance (κ = 0.465; p < 0.001), indicating that both methods are capable of identifying the presence of nutritional risk, although they are not fully concordant. Srinivasaraghavan et al.36 assessed the concordance of CC with nutritional screening methods and showed that CC exhibited good concordance, which could be used to identify outpatient malnutrition and facilitate early nutritional intervention, corroborating our findings. Other studies have also explored the applicability of CC as a substitute method for indicating sarcopenia, malnutrition, and the risk of hospital readmission, with some of them covering populations from middle-aged adults to older adults individuals37,38.
It is important to emphasize the significance of nutritional risk assessment, given the high prevalence of malnutrition at hospital admission and the negative outcomes associated with this clinical condition14,31. Early identification of nutritional risk allows for timely therapeutic interventions, such as nutritional therapy, to restore or maintain the nutritional status of patients with HNC, aiming to improve clinical outcomes and quality of life. CC facilitates early detection of risk and nutritional status and has shown good performance as a substitute method to predict nutritional risk, making it useful when other screening methods are not applicable.
This study has some limitations. Data were obtained from a single reference oncological center, and since this is a cross-sectional study, it was not possible to track future outcomes of the evaluated patients or establish causal relationships. Furthermore, although the number of participants met the estimated value from the sample calculation, the selection was made by adherence rather than probabilistic sampling. Patients with HPV-positive status were not excluded from the sample, which may have influenced the epidemiological profile of the sample.
Our findings complement the current literature and reinforce the importance of rigorous nutritional risk monitoring upon the admission of oncological patients, particularly in the context of HNC, since factors associated with the patient profile are closely linked to nutritional risk.
CONCLUSION
The results of this study highlight the relevance of assessing nutritional risk in patients with HNC, particularly in a context of high prevalence of malnutrition at the time of diagnosis. Factors such as low income and CC are significantly associated with this risk.
In addition to determining the factors associated with nutritional risk, it was found that the correlation and concordance between the NRS-2002 and CC reinforce the utility of CC as an additional method to identify nutritional risk, as a smaller CC is associated with nutritional risk in the evaluated population.
