In a relatively large cohort of 310 patients with locoregionally advanced HNSCC compared with previously published studies7,9,20−23, we retrospectively analyzed whether omitting a PEG compared to prophylactic PEG insertion is associated with an increased risk of complications leading to a TRUH. Although the institutional policy was to offer pPEG to all patients with locoregionally advanced HNSCC, physicians were less keen on insisting that patients with a possibly lower risk profile should receive a pPEG. Moreover, some patients refused the pPEG regardless of their risk profile. Therefore, pPEG placement tended to be used more frequently in patients with a higher risk profile and worse prognosis (comprising general condition, tumor size, age, hypopharyngeal tumor localization; Table 1). In the nPEG group, apart from (chemo)-RT-related side effects, dysphagia/dehydration/malnutrition (n = 8; 20%) was the most frequent cause of TRUH, whereas in the pPEG group, apart from (chemo)-RT-related side effects, PEG-related complications frequently led to TRUH (n = 11; 14%). PEG tube placement is associated with the risk of complications; however, there is a great deal of variability in the reported incidence of such complications11,20,24−29. The difference in the incidences of complications is partly due to the various definitions and populations analyzed. For example, complications are more likely to occur in older patients with comorbidities, especially those with an infection or history of aspiration30. Compared with the publication of Silander et al.10, our rate of PEG-related complications is relatively high (14% versus 1%); however, it is relatively low compared with a prospective study reporting complication rates at 2 weeks and 2 months (39% and 27%, respectively)24. We hypothesize that our pPEG cohort is a different, more fragile patient population that tends to have more complications compared with that studied by Silander et al.10 and our nPEG population. Furthermore, we suspect that patients – like those studied by our Swedish colleagues10 – who are willing to be included and randomized in a study, are more compliant than the patients with HNSCC seen in our everyday practice, over two-thirds of whom have a positive history of alcohol abuse and more than 87% a positive history of tobacco use31. Patients with severe clinical and psychosocial impairment and fewer economic resources are more likely to experience treatment compliance problems31. There is an increasing incidence of oropharyngeal cancer, especially in younger patients, and a decrease in the previously known risk factors for HNSCC of smoking and alcohol use32,33. Previously, typical patients with HNSCC tended to be heavy drinkers or smokers; however, human papillomavirus (HPV)-associated HNSCC in younger, fitter, and possibly more compliant patients increasingly represent the majority of at least oropharyngeal disease34. This interesting aspect should be kept in mind before considering that in our entire cohort, up to one-fifth of patients had a TRUH besides the (chemo)-RT-induced TRUH – due to dysphagia/dehydration/malnutrition (20%) in the nPEG group or postoperatively after PEG insertion (12%) in the pPEG group. The physician and patient have to face the additional risks associated with an invasive procedure, such as PEG tube placement, or those arising from not performing a supportive surgical procedure to allow sufficient oral intake, such as dysphagia/dehydration/malnutrition.
Further differences between our cohort and the Swedish study10,11 can be seen with regard to weight loss, BMI, and OS between the pPEG and nPEG groups. The increased weight loss and BMI differences during RT in the nPEG versus pPEG groups could be explained not only by the greater compliance of patients but also by the prospective setting – and therefore thorough monitoring by nutrition counselors in the nPEG cohort – of the Silander et al. trial10. Nutrition counselors were not systematically involved in the treatment of our patients, and some patients categorically refused nutrition counseling. The higher risk profile in the pPEG group more easily explains the OS difference versus the nPEG group (tumor size, age, tumor localization; Table 1), as OS is known to be worse in patients with larger primary tumors and hypopharyngeal tumor35–38. Other limitations of our study, apart from the different risk profiles of the nPEG and pPEG groups, include its retrospective nature and the lack of stratification according to HPV status.
With future changes in the HNSCC population, therapy regimens, and side-effect profile according to the HPV status, further analyses of the indication for a PEG is necessary39.