Letter to the editor regarding: “Development of the ProPal-COPD tool to identify patients with COPD for proactive palliative care”

Dear editor

We read with interest the recent article by Duenk et al entitled “Development of the ProPal-COPD tool to identify patients with COPD for proactive palliative care” recently published in the International Journal of Chronic Obstructive Pulmonary Disease.¹ We fully agree with the authors on the capital importance of proactive palliative care (PPC) in COPD, as well as its underutilization. Any article that contributes to increasing knowledge of PCC and its use with COPD patients who can benefit from it is to be celebrated. We also agree with the authors on the unpredictability of prognosis in COPD, especially the uncertain evolution of severe exacerbations and the impact of comorbidities.

We believe that the article deserves some considerations. First and most important in our opinion, PPC should not be limited to the terminal phase of chronic diseases such as COPD; it can be delivered alongside standard therapies, according to the needs and preferences of patients regardless of the risk of death in the short or medium term. Obviously, this does not mean that prognostic prediction scales are not useful in prognosis prediction. Nevertheless, its exclusive use may exclude many patients who could benefit from PPC.² For instance, in the model proposed by Duenk et al, seven dichotomic predictor variables were suggested to consider PPC based on their relationship with 1-year mortality. However, it seems clear that a COPD patient with severe airflow obstruction, disabling dyspnea, and several previous hospitalizations, although strictly not meeting all the recommended criteria, is a candidate for PPC. In this patient, PCC includes advanced care planning conversations, with their perspectives in case of a poor evolution in future exacerbations and preferences concerning the ceiling of treatment such as cardiopulmonary resuscitation, admission to the intensive care unit, and invasive mechanical ventilation. Additionally, pharmacological and non-pharmacological measures for dyspnea or others symptom treatment should be considered and discussed with the patient. In this example, the onset of these measures should be independent of other predictor variables and of foreseeable life expectancy. Of note, PPC measures by themselves do not shorten life but may be associated with increased survival.³

Second, from a strictly methodological point of view, we cannot forget that the model is built from just 30 positive patients. Even by using most sophisticated statistical procedures, we still have only little information about the general behavior of this population, and moreover, this is a strong limitation due to the relevance of the topic and the final decisions taken. Additionally, all prognosis models must be validated in a different external cohort, to avoid the risk of overestimation inherent in development cohorts. In the present study, only internal validation was performed. Finally, the exclusive use of the receiver operating characteristic curves and area under the curve for 1-year mortality limits consideration to the vital status of the patient 1 year after discharge, regardless of the time of death, and clearly survival time is relevant in this population. Cox regression analysis and Kaplan–Meier curves in patients with and without ProPal-COPD criteria would be of help to further clarify the result.