Introduction
Gastrostomy is commonly used as an enteral feeding route to supplement nutrition in patients with head and neck cancer. Gastrostomy tube use has increased since the publication of several systematic reviews demonstrating its safety.Reference Grant, Bradley, Pothier, Bailey, Caldera and Baldwin1, Reference Wollman, D'Agostino, Walus-Wigle, Easter and Beale2 However, the indications for its use remain poorly defined, and variable practices are employed across head and neck centres.Reference Paleri and Patterson3
Several studies have examined the role of different clinical and patient factors that predict the need for gastrostomy tube placement.Reference Nugent, Parker and McIntyre4–Reference Moor, Patterson, Kelly and Paleri7 However, there are no published data on factors affecting the duration of gastrostomy tube retention. Interviews of disease-free patients who retain their gastrostomy tube for over 12 months have identified several clinical, social and personal factors,Reference Owen, Patterson, Johnson and Paleri8 many of which are difficult to quantify.
This single-centre, retrospective study aimed to analyse the quantifiable patient-, tumour- and treatment-related factors that may influence the duration of gastrostomy tube retention in patients who are disease-free after treatment for head and neck cancer.
Materials and methods
Study population
Patients with mucosal tumours of the upper aerodigestive tract who underwent gastrostomy tube placement between January 2003 and December 2007 were identified via the dietetic service and included in this retrospective study. Patients who underwent gastrostomy tube placement for supportive care alone or for disease recurrence were excluded, as were those for whom data were insufficient.
Clinical variables
The clinical indices recorded were: tumour–node–metastasis (TNM) stage, tumour stage, co-morbidity grade, tumour site, treatment modality and deprivation score. The only data collected from the notes were those aspects pertaining to the duration of gastrostomy tube retention. All other patient demographic and clinical data were identified from a local head and neck oncology database.
The demographic data collected for this study included postcode and age. The primary outcome measure was the duration of gastrostomy tube retention (in days). The sixth edition of the TNM Classification of Malignant Tumours was used for staging purposes.Reference Sobin and Wittekind9 The treatment modalities were categorised into unimodality (surgery only or radiation only) or multimodality (surgery and radiotherapy, or concurrent chemoradiation).
Patients were divided into deprivation quintiles on the basis of their postcode, according to their Townsend score.10, Reference Townsend, Phillimore and Beattie11 The Townsend index is based on four variables from the 2001 Census: unemployment, overcrowding, non-home ownership and non-car ownership. The score is derived from the sum of standardised scores for each variable, and an increased score represents increased deprivation. The Townsend score is categorised into quintiles: 1 represents the most affluent and 5 represents the most deprived.
Co-morbidities are diseases that coexist with the patient's head and neck cancer. The burden of co-morbidity was estimated using the Adult Comorbidity Evaluation 27-item index (ACE-27). This index, which has been validated in head and neck cancer, categorises co-morbidity into grades of severity: none (0), mild (1), moderate (2) and severe (3).Reference Piccirillo, Costas, Claybour, Borah, Grove and Jeffe12
Statistical analysis
Univariate analyses of gastrostomy tube retention were restricted to disease-free survivors in order to eliminate competing variables in those patients with uncontrolled disease.
Clinical variables were dichotomised for statistical comparison: (1) co-morbidity grade, none or mild versus moderate or severe; (2) TNM stage, stages I–III versus stage IV; (3) tumour stage, stages T0–2 versus stages T3–4; (4) tumour location, oral cavity and larynx versus pharynx and unknown primary; (5) treatment modality, multimodality versus single modality; and (6) Townsend deprivation quintile, 5 versus 1–4. All analyses were performed using the Statistical Package for the Social Sciences software, version 17 (SPSS; Chicago, Illinois, USA).
The categorisation of tumour location into two groups for analysis was considered pragmatic. This is in line with the treatment policy at our centre: the treatment fields for tumours of the pharynx and unknown primary sites are wider, and these patients tend to have greater swallowing-related morbidity. Given the numbers of patients with higher deprivation (quintile group 5) in this cohort, Townsend quintile groups 1–4 were grouped into one category.
Results
Of 151 patients who underwent gastrostomy tube placement between January 2003 and December 2007, we identified 132 whose clinical records were considered to have sufficient data for inclusion in this study. The demographic and clinical characteristics are shown in Table I.
Table I Patients' demographic and clinical characteristics
TNM = tumour–node–metastasis
The mean patient age was 63 years (standard deviation = 10.97, range of 27 to 89 years). The mean duration of gastrostomy tube retention in all 132 patients was 16.75 months (standard error of the mean (SEM) ± 1.36, median = 12.63). The duration of gastrostomy tube retention was longer in the 66 survivors (mean = 21.26 months, SEM ± 2.22, median = 15.9). The 66 patients with uncontrolled disease and those who died had a lower mean duration of 11.89 months (SEM ± 1.46, median = 8.25).
The gastrostomy tube retention data were analysed further in survivors. Univariate analysis was performed using the independent samples t-test and adjusted for unequal standard deviations. This revealed that patients with a co-morbidity grade of 2 or 3 retained their gastrostomy tubes for significantly longer (mean duration of 29.7 months); those with a co-morbidity grade of 0 or 1 had a mean duration of 18.1 months (p = 0.041). The univariate analysis identified no other significant variable. The three most common categories of co-morbidity were: cardiovascular disease (52 per cent), chronic obstructive pulmonary disease (COPD) (26 per cent) and substance abuse (29 per cent) where alcohol was the only substance abused. Thirty-eight per cent of survivors (25 of 66) who underwent gastrostomy tube placement had a co-morbidity grade of 0. The durations of gastrostomy tube retention by group are shown in Table II.
Table II Clinical parameters and gastrostomy tube retention
*Independent samples t-test, two-tailed. †Survivors only. Pts = patients; GT = gastrostomy tube; mth = months; TNM = tumour–node–metastasis
A significant proportion of our cohort came from the most deprived areas as defined by Townsend deprivation quintiles. Although the mean duration of gastrostomy tube retention was longer in patients living in the Townsend quintile 5 regions (24.5 months) compared with those living in less deprived areas (18.5 months), this difference was not statistically significant.
Discussion
In our retrospective study of 132 head and neck cancer patients, a higher co-morbidity grade was significantly associated with a longer duration of gastrostomy tube retention in survivors. The data also demonstrated that patients living in areas of greater deprivation retained their gastrostomy tube for longer; however, this finding was not statistically significant.
Enteral feeding by gastrostomy tube is widely used in the nutritional management of patients receiving treatment for upper aerodigestive tract tumours. While several studies have looked at the clinical and treatment factors that influence the decision to place a gastrostomy tube, this is the first study to analyse the impact of quantifiable demographic, patient, tumour and treatment factors on the duration of gastrostomy tube retention.
The variables chosen for analysis were based on what we clinically judged may have an impact on the duration of gastrostomy tube retention. Given the cohort, we determined that if a multivariate analysis was required, the use of more than six variables would undermine the strength of any conclusions.
Our results demonstrated that increased co-morbidity grade was the only independent prognostic factor for gastrostomy tube dependence. Given the high rate of smoking and alcohol use often seen in patients with head and neck cancer, it is not unexpected that cardiovascular disease, COPD and alcohol abuse were the most common co-morbidities in our cohort. These observations therefore support in part a study of head and neck cancer patients that found alcohol dependence to be significantly associated with long-term gastrostomy tube use.Reference Schweinfurth, Boger and Feustel13
Our recent review confirmed the lack of consensus in using primary site alone as an indication for gastrostomy tube placement.Reference Paleri and Patterson3 Various other factors have also been implicated in the decision making for gastrostomy tube placement. A study investigating factors independently associated with feeding tube placement identified the following variables: oropharynx or hypopharynx tumour site, stage III or IV, disease flap reconstruction, current tracheostomy, chemotherapy, and increased age.Reference Cheng, Terrell, Bradford, Ronis, Fowler and Prince14 Surgery to the tongue base and pharynx, and flap reconstruction,Reference Schweinfurth, Boger and Feustel13 and radiation dose delivered to the inferior constrictors and the cricopharyngeal inlet,Reference Li, Li, Lau, Farwell, Luu and Rocke15 have also been cited as factors that predispose a patient to prolonged enteral support.
Some of those factors were analysed in the current study; however, none were correlated with the duration of gastrostomy tube retention. It has been postulated that a lack of social support and a higher index of social deprivation may play a role in the duration of gastrostomy tube use. Our data showed an increase in gastrostomy tube retention in relation to deprivation index, but the difference was not statistically significant; however, this finding is limited by the lack of patients in our cohort from less deprived areas. None of the other factors analysed in this study were associated with duration of gastrostomy tube retention.
Advanced tumour stages and prophylactic gastrostomy tube placement have both been associated with gastrostomy tube retention at 12 months.Reference Chapuy, Annino, Snavely, Li, Tishler and Norris16, Reference Chen, Li, Lau, Farwell, Luu and Stuart17 Our study found no significant relationship between gastrostomy tube retention and tumour stage. Of note, we employ a highly selective policy in offering patients gastrostomy tube placement, which may account for the longer mean duration of gastrostomy tube retention compared with other publications.Reference Oates, Clark, Read, Reeves, Gao and Jackson18–Reference Wiggenraad, Flierman, Goossens, Brand, Verschuur and Croll20 The decision to offer gastrostomy tube placement is made by a multidisciplinary team, following nutritional and swallowing assessments. Factors such as patient preference, social support and expected treatment-related dysphagia are also taken into account.
The presence of a feeding tubeReference Terrell, Ronis, Fowler, Bradford, Chepeha and Prince21 and gastrostomy tube retention at one yearReference El-Deiry, Futran, McDowell, Weymuller and Yueh22 are significantly associated with a decreased quality of life (QoL). Importantly, pre-treatment QoL and co-morbidity are also predictive of QoL at one year post-treatment.Reference El-Deiry, Futran, McDowell, Weymuller and Yueh22 Given the significant negative associations between gastrostomy tube use, co-morbidity and QoL, it is vital that further prospective studies are conducted to examine the factors affecting gastrostomy tube retention. Gastrostomy tube retention is likely to be affected by many factors, but a clear understanding of these factors will provide valuable insight for treatment planning and information given to patients.
• Head and neck cancer patients often require nutritional support via a gastrostomy tube
• Patients remain gastrostomy tube fed for varying durations, the reasons for which are poorly understood
• This is the first study to analyse the impact of quantifiable demographic, patient, tumour and treatment factors on the duration of gastrostomy tube retention
• This study demonstrated a correlation between higher co-morbidity grade and longer duration of gastrostomy tube retention
The findings of this study highlighted co-morbidity as an important factor which significantly affected the duration of gastrostomy tube retention. The study is, however, limited by the retrospective collection of data and the need to dichotomise data for analysis. Nevertheless, the association between co-morbidity and longer duration of gastrostomy tube dependence may be taken into account when planning treatment or counselling patients about their long-term function following cancer treatment. Co-morbidity should be considered as a variable in future prospective studies of gastrostomy tube use.
Conclusion
In this study, TNM stage, tumour stage, primary site, treatment modality and deprivation were not independently predictive factors for the duration of gastrostomy tube retention in patients with upper aerodigestive tract tumours. However, increased co-morbidity grade was significantly associated with increased duration of gastrostomy tube retention. We postulate that gastrostomy tube retention is likely to be affected by many factors, with few single variables being independently important. The role of other factors influencing prolonged gastrostomy tube dependence remains speculative. Despite the limitations of this retrospective study, the findings add weight to the importance of acknowledging co-morbid disease when planning head and neck cancer treatment. We welcome future prospective clinical studies to verify co-morbidity as an independent predictive factor for gastrostomy tube retention.
