An exploratory study to identify risk factors for the development of capecitabine-induced Palmar Plantar Erythrodysesthesia (PPE)
Date
Authors
Advisors
Journal Title
Journal ISSN
ISSN
DOI
Volume Title
Publisher
Type
Peer reviewed
Abstract
Background Previous literature showed contradictory evidence on the subject of predictors of chemotherapy-induced Palmar Plantar Erythrodysesthesia (PPE). While there is evidence to suggest that dose and schedule of the drugs play a large role, the fact that many still go on to develop severe PPE following dose reduction would indicate that there are other factors involved. Since the incidence of PPE is more prevalent during the first three cycles of treatment this would also indicate that there are factors other than a cumulative effect. The contradictory evidence in the literature relates to biographical factors, performance status, co-morbidities and renal function. There is a lack of empirical evidence to support the theory that PPE is caused by damage to the microcapillaries due to everyday activities that cause friction or pressure to the hands or feet.
Purpose The aim of this exploratory study was to identify pre-treatment risk factors for the development of PPE prior to cycle four.
Patients and methods The study was made up of two phases, a retrospective phase and a prospective phase, using mixed strategies to collect data. Thus providing two independent samples to compare and validate or refute results. Phase I A retrospective notes review of patients who had received Infusional 5FU or capecitabine containing regimes over a 1 year period (n=392). Phase II Prospective data collection from participants receiving capecitabine monotherapy (n = 125). Data was collected during semi-structured interviews, from participant’s diaries, physical examination of the hands and feet and notes review. Data relating to activities that cause friction, pressure or heat were collected during this phase.
Data from both samples were analysed independently using bivariate (chi-square and t-test) tests where each independent variable was analysed against PPE. The variables which achieved statistical significance were entered into a multivariate (binary logistic regression) model. The multivariate analysis employed a specific modelling algorithm using a relaxed alpha value applied to various entry methods to produce multiple models. The outcomes from these models were entered into a ROC curve test to establish which model was the best predictor of PPE.
Results Phase I The bivariate analysis demonstrated that those at most risk of developing PPE prior to cycle 4 of capecitabine monotherapy were males with non-metastatic colorectal cancer, who had either developed PPE with previous chemotherapy regimes or not had previous chemotherapy and who started their treatment during the winter months. When variables were combined in a multivariate logistic regression model, those that were associated with an increased risk of PPE were male, no metastatic spread, no inflammatory condition as co morbidity, smoked, did not drink, had weight loss prior to treatment, a low/normal pre treatment ALP level and started their treatment during the winter.
Phase II The bivariate analysis demonstrated that those at most risk of developing PPE prior to cycle 4 of capecitabine monotherapy were those with no metastatic disease, had an inflammatory condition as co morbidity, were receiving capecitabine as adjuvant treatment, had a good performance status (0-1) and had a tendency to have warm hands. When variables were combined in a multivariate logistic regression model, those that were associated with an increased risk of PPE were younger (< 65) had no metastatic disease, an inflammatory condition as co morbidity, drank alcohol regularly, had a good performance status, had not received previous radiotherapy, were overweight or obese, had a pre treatment creatinine clearance of 30-50mls/min and had a tendency to have warm hands.
Conclusions Similarly to the literature, contradictory findings were seen between the two samples within this study. There was only one variable that was associated with the development of PPE prior to cycle 4, which was the absence of metastatic disease. Limitations of retrospective data may explain variation in some variables which may have been underreported; however it is likely that it is not possible to identify specific factors that increase the risk of PPE. This is the first study to have collected and analysed data related to friction, pressure and heat causing activities. These activities have been suggested as increasing the risk of developing PPE and form the basis of patient education to avoid these activities. Data from this study indicates that only a tendency to have warm hands is associated with an increased risk of PPE. Whilst this finding would need validating in larger studies, it is a unique contribution to the body of knowledge of PPE. This finding indicates that avoidance of activities that cause friction and pressure has no evidence base. Patients may therefore be avoiding activities that add to their enjoyment which at this stressful time in their lives may add to any psychological distress. Despite limitations of this study, the importance of the findings presented here lie in its usefulness in shaping future research to investigate identified variables, where before no direction was available.