Page 198 - Functional impairment and cues for rehabilitation of head and neck cancer patients -

196
Chapter 8
Figure 2 Directed acyclic graph with confounders and mediators in the relation between sarcopenia and prolonged feeding tube dependency. Arrows indicate direction of effect. Dotted arrows indicate direction of effect for unavailable confounder (health literacy). BMI, body mass index; FOIS, functional oral intake scale; SES, socioeconomic status.
Cut-off for sarcopenia
Since no normal values of the neck SMI exist, the optimal cut-off value of the neck SMI for predicting prolonged tube feeding was determined using the Youden point of the receiver operating characteristic (ROC) curve of neck SMI versus prolonged feeding tube dependency. To obtain an indication for clinical usefulness of pretreatment neck SMI measurements, the number of patients below this cut-off value (indicating sarcopenia) was assessed, stratified by their (predicted) probability on prolonged (> 90 days) feeding tube dependency according to our previously published prediction model (23). This model included the clinical predictors FOIS, BMI, weight loss and T classification and had the following formula: Y = 0.617 + (0.145∙T2 + 0.382∙T3 + 0.727∙T4) + (-0.067∙BMI) + (0.543∙FOIS < 7) + (0.356∙weight loss < 10% + 0.980∙weight loss > 10%).