Impact of Palliative Care in Evaluating and Relieving Symptoms in Patients with Advanced Cancer. The DEMETRA Study

Background: Cancer patients experience a number of symptoms throughout the course of the disease. We aimed to provide a comprehensive analysis of the symptom burden in patients with advanced cancer at admission to specialist palliative care (PC) services and seven days later, to estimate the immediate impact of PC intervention. Patient and methods: The analysis was based on an observational, prospective, multicenter study (named DEMETRA) conducted in Italy to outline the prole of patients, families and PC services in different care settings (hospital, hospice and home care). The prevalence and intensity of symptoms were assessed using three tools, including the Edmonton Symptom Assessment System (ESAS). Results: Five PC centers recruited 865 cancer patients. Thirty-three different symptoms were observed at baseline, the most frequent being asthenia (85%) and lack of appetite (71%). Two-thirds of patients experienced six to twelve simultaneous symptoms. The intensity of the most frequent symptoms according to ESAS varied from 5.5 for asthenia to 3.9 for nausea. The presence and intensity of physical symptoms increased with increasing levels of anxiety and depression. After seven days, prevalence decreased signicantly only for nausea and breathlessness, while intensity diminished signicantly for almost all symptoms. At admission we noted a correlation between patients' symptoms and the care setting. After one week, the symptom intensity was uniformly reduced in all settings. Conclusions: The study conrmed the considerable symptom burden of patients with advanced cancer. PC intervention signicantly lessened the severity of symptoms, despite the patients’ advanced disease and short survival.


Introduction
Palliative care (PC) is mainly addressed to the evaluation and treatment of physical, functional, psychological, social, and spiritual symptoms throughout the course of chronic and progressive diseases, especially in the advanced and terminal phases [1,2], with the aim of improving, or at least preserving, patients' quality of life.
Most evidence in recent decades has been related to the treatment of symptoms in cancer patients, at any stage of the disease and of whatever causal origin. In fact, symptoms may derive from the primary disease or its secondary localizations, but also from cancer therapies and even from treatments aimed to control comorbidities or other symptoms such as pain [3][4][5][6][7][8][9]. Nausea and vomiting, for instance, may be due to the localization of the tumor, to impaired gastric emptying or delayed gastrointestinal transit, to antineoplastic agents or radiation therapy, or to opioids given to relieve pain [10,11]. In the patient's evaluation the distinction of the symptom's causal factors may be irrelevant, but for the physician understanding the pathogenetic mechanisms is fundamental to setting up speci c therapies according to the cause. Moreover, the prevalence and severity of symptoms vary across the phases of illness [12].
Pain, nausea and depression have been reported to be relatively stable over the last six months of life, while dyspnea, drowsiness, poor well-being, lack of appetite and asthenia tend to increase in severity over time. In particular, the patient's symptom burden changes during the last days of life, when PC becomes the speci c and often only therapeutic approach [13]. The management of this clinical picture -with some symptoms tending to disappear and others taking their place -is a central and crucial issue in controlling patients' suffering.
In previous studies, only selected aspects of these symptoms were considered. We therefore carried out this analysis with the objective of providing a comprehensive evaluation of the symptom burden in patients with advanced cancer from the moment of admission to a PC service. We estimated the prevalence and intensity of symptoms, the number of symptoms experienced simultaneously, their prevalence based on the primary tumor site, and the way in which the presence and severity of anxiety and depression could modify the physical symptoms. We reassessed symptom prevalence and intensity after seven days to estimate the immediate impact of PC, paying attention also to the in uence of the care setting on the clinical expression of symptoms.

Study design
The data of this analysis derive from an observational, prospective, multicenter study (named DEMETRA) aimed at outlining the pro le of patients, families, care services and clinical features from the moment of admission to PC centers. The general presentation of DEMETRA has been published elsewhere [14].
Given the considerable extension of the information collected and the results achieved, the Scienti c Committee decided to divide its contents into separate publications, each deepening a speci c theme.
Five Italian PC units based in Florence, Forlì, Lecco, Palermo and Rome participated in the study, which comprised a total of 1013 individuals suffering from various advanced chronic diseases. Most were cancer patients (85%), treated in three different care settings: home, hospice and hospital. The analysis of symptoms reported here concerns only cancer patients, given the high prevalence of cancer among the population studied and the speci city of the clinical conditions and symptoms associated with cancer compared to other advanced diseases.
The inclusion criteria for the DEMETRA study were age ≥ 18 years; presence of a chronic and progressive illness requiring palliative intervention; and written informed consent for study participation and personal data processing. Patients who were already in the care of a PC network and those who could not ensure regular follow-up were excluded.
Eligible patients were recruited over a six-month period (May to November 2017). Each patient was followed up for 12 months, until November 2018.

Data collection
Among the patient information collected in the DEMETRA study were sociodemographic data, primary tumor site, presence and sites of metastasis, number of concomitant diseases, care setting (hospice, home care or hospital) and symptoms [14]. Data were collected at the baseline visit (day 0) and after seven days (day 7), to assess the rst impact of palliative treatments.
Symptoms were assessed using three tools: 1) the interRAI-PC, an instrument evaluating a wide range of PC issues, including items related to symptoms [15]; 2) a supplementary list of items to detect symptoms and clinical aspects not covered by the interRAI-PC including rattle, dysphagia, dysgeusia, neuropathic pain, bowel occlusion, liver and kidney functional impairment, paresis, motor disorders, spinal cord compression, pleural effusion and intracranial hypertension; 3) the Edmonton Symptom Assessment System (ESAS), which estimates the prevalence of nine main symptoms (pain, asthenia, nausea, depression, anxiety, drowsiness, breathlessness, poor well-being, lack of appetite) and their intensity, measured with a numeric rating scale from 0 to 10 [16,17].

Statistical analysis
Categorical variables were described by absolute and relatives frequencies, while continuous variables were summarized using mean values with the corresponding standard deviation (SD). The differences in the prevalence of symptoms across levels of anxiety and depression were evaluated using Pearson's chisquared test, while the Kruskal-Wallis rank-sum test was used to evaluate differences in mean symptom intensity between different levels of anxiety and depression. McNemar's chi-squared test was used to evaluate differences in the prevalence of symptoms between days 0 and 7. To compare symptom intensity at day 0 and day 7, we applied the paired-samples Wilcoxon test. All data were analyzed using the R software package (version 3.6.1/2019).

Results
The owchart of patients evaluated in the present study is shown in Fig. 1. From a total of 1013 patients, we selected 865 cancer patients. Of these, 780 answered the ESAS questionnaire and were considered for the intensity of symptoms at baseline, while 508 (65%) were also evaluated at day 7.
The baseline characteristics of patients are summarized in Table 1. Fifty-two percent were women and the mean age was 74 years. Fifty-ve percent of patients were assisted at home. The most frequent primary tumor sites were the digestive system, lung, urinary and reproductive system. About 80% of patients had metastases, the most frequent sites being liver, lung and bone. SD, standard deviation.  Table 3 shows the prevalence of the main symptoms observed at baseline according to the primary tumor site. Breathlessness was mainly frequent in lung cancer, pain in pancreatic cancer, anxiety and drowsiness in breast cancer, and lack of appetite in colorectal cancer. Nine of the 10 most frequent symptoms were included in the ESAS questionnaire. This allowed us to measure the intensity of these symptoms in the whole population and in symptomatic patients, as shown in Table 3. In the rst group, the intensity ranged from 5.2 (asthenia) to 1.7 (nausea); in the second, from 5.5 (asthenia) to 3.9 (nausea). The prevalence and intensity of the four main physical symptoms (pain, breathlessness, nausea and asthenia) signi cantly increased with increasing levels of anxiety and depression (Supplementary Table 4). SD, standard deviation.
In Table 4, the prevalence and intensity of symptoms are compared in patients evaluable at both visits, on day 0 and day 7. After one week there was a signi cant decrease in the prevalence of breathlessness and nausea only, but the intensity decreased signi cantly for all symptoms except anxiety, depression and drowsiness. Among patients with severe symptom intensity at baseline (score ≥ 6) we observed a reduction in intensity for all symptoms (Supplementary Table 5). After one week, breathlessness was signi cantly reduced in patients treated at home, while nausea was signi cantly reduced in those treated at home or in hospice (Supplementary Table 6). Anxiety and depression signi cantly increased in those treated at home and in hospital, respectively. Pain signi cantly decreased in all care settings; anxiety only in patients treated in hospital; breathlessness, poor well-being, lack of appetite, nausea and asthenia in those treated at home; and lack of appetite, nausea and asthenia in those in hospital (Table 5).

Discussion
This study evaluated the symptom burden in a cancer patient population during the last days of life, at the time of admission to specialist PC. The decision to appraise only cancer patients was prompted by the need for a clinically homogeneous population but also by the high proportion of cancer patients among those included in the DEMETRA study. In addition, cancer patients' average life expectancy is generally short at the time of admission to PC. The median survival in our population was 29 days, and only about 65% of patients were evaluable after the rst seven days. Due to the advanced phase of disease, PC health professionals had little time to reach therapeutic goals, which were mainly to alleviate symptoms. We sought to evaluate whether and how the symptom pro le of patients could be changed in a few days as a result of PC intervention.
We observed the presence of many symptoms at the time of PC admission. Asthenia was the most frequent, followed by poor well-being, lack of appetite, drowsiness, pain, depression, constipation, anxiety, breathlessness and sleep disorders, which were present in 40-70% of patients. The large number of symptoms is characteristic of the advanced stage of cancer, as previously documented [18].
Furthermore, patients experienced a large number of symptoms simultaneously, as also previously observed [9,19]. This is an important and underestimated clinical problem. For patients the multitude of symptoms is very distressing, while for physicians treating all the symptoms is complex due to the number of drugs needed, their potential toxicity, and the risk of drug-to-drug interactions. This makes it necessary to decide which symptoms should be primarily treated, with which drugs, and what results are expected, important aspects that open a window on future insights. In our study, frequent combinations of simultaneous symptoms nearly always included asthenia, lack of appetite, drowsiness and poor wellbeing, consistent with the ndings of others [20,21]. Such combinations entail a negative physical and psychological condition, with lack of energy and interests and detachment from life.
The prevalence of symptoms at baseline differed according to the site of the primary tumor, as also reported elsewhere [22,23]. The association between speci c tumor sites and cancer symptoms is clear, even at advanced stages of the disease when widespread metastasis can dilute this speci city. Instead, some symptoms, such as pain and asthenia, tend to occur in all tumors.
The intensity of symptoms ranged from 5.2 for asthenia to 1.7 for nausea in the overall population, and from 5.5 for asthenia to 3.9 for nausea in symptomatic patients, showing analogies to earlier observations [24,25]. Interestingly, the prevalence and intensity of physical symptoms such as pain, breathlessness, nausea and asthenia correlated with the presence and severity of anxiety and depression, re ecting the integration between physical and mental suffering.
We also evaluated whether and to what extent PC treatments could change the symptom pro le of patients after one week. A week is a short period of time, but not for patients with terminal cancer. Their life expectancy is very limited, and as a consequence every therapeutic intervention must be immediately effective. The results obtained in one week are therefore an important measure of the clinical impact of PC. In our study, the prevalence of symptoms between day 0 and 7 was unchanged except for breathlessness and nausea, which tended to decrease signi cantly. Considering that the clinical situation quickly worsens in advanced cancer, even the containment of a few symptoms should be seen as positive. Conversely, symptom intensity was signi cantly reduced after one week of treatment, with the exception of depression, anxiety and drowsiness. This was an important clinical result, indicating the relief of suffering at such a highly critical time.
Finally, changes in the prevalence and intensity of symptoms between day 0 and 7 were evaluated in the three settings of care: home, hospice and hospital. Some symptoms had a different initial prevalence depending on the setting. Poor well-being, pain and nausea were more frequent in patients at home, while lack of appetite and depression prevailed in hospice patients. After one week, the reduction of intensity was almost uniform for all symptoms, with negligible differences related to the setting of care.
The main strength of this study is that it provides a comprehensive analysis of various aspects of the symptoms experienced by cancer patients in the nal period of life. Its main limitation is the impossibility of describing the treatments given for such symptoms and the subsequent clinical results, because the DEMETRA study protocol did not include the collection of these data.

Conclusions
The primary aim of this study was to de ne the clinical symptom pro le of 865 patients with advanced cancer from the moment they accessed PC services. The study con rmed that the number and severity of symptoms in these patients was substantial. Multiple symptoms were present and many were perceived as severe. Our results make it clear that PC intervention can signi cantly lessen the severity of symptoms, even at the advanced stages of cancer. This further con rms the need for these patients to receive competent PC. In our opinion, such competence concerns not only PC specialists but also oncologists, general practitioners and any other physicians dealing with patients at an advanced stage of disease.

Declarations
Ethics approval and consent to participate: The DEMETRA study was approved by the ASST Lecco Ethics Committee on 1st December 2016 and subsequently by the institutional review boards of each participating center. Written informed consent for participation in the study and processing of personal data was collected from all recruited patients before any study-related activities were carried out.  Figure 1