Palliative Performance Status, Heart Rate and Respiratory Rate as Predictive Factors of Survival Time in Terminally Ill Cancer Patients

Transcription

1 Vol. 31 No. 6 June 2006 Journal of Pain and Symptom Management 485 Original Article Palliative Performance Status, Heart Rate and Respiratory Rate as Predictive Factors of Survival Time in Terminally Ill Cancer Patients Cristina de Miguel Sánchez, MD, Sofía Garrido Elustondo, MD, Alicia Estirado, MD, Fernando Vicente Sánchez, MD, Cristina García de la Rasilla Cooper, MD, Andrés López Romero, MD, Angel Otero, MD, and Luis García Olmos, MD Home Care Support Team (C.M.S., F.V.S.), Madrid s Area 7; Investigation Unit (S.G.E.), Madrid s Area 7; Madrid s Area 7 (A.E., C.G.R.C.); Primary Care Administration (A.L.R.), Madrid s Area 3; Preventive Medicine and Public Health Department (A.O.), Universidad Autónoma Madrid; and Primary Care Research (L.G.O.), Madrid s Area 2, Madrid, Spain Abstract To determine which symptoms, signs, and characteristics that define the patient s functional status predict the survival time in terminally ill cancer patients, a prospective longitudinal study was conducted with terminally ill cancer patients followed by a Home Care Support Team. Patients were followed up with at least weekly visits until death, collecting variables at each visit. A Cox multivariate regression analysis took into account all the follow-ups in the same patient. Ninety-eight patients were studied, and 250 evaluations were done. The mean age was 72 years. The median survival was 32 days. In the multivariate analysis, three independent variables were identified: Palliative Performance Score of 50 or under, heart rate of 100/minute or more, and respiratory rate of 24/minute or more. The variables that were found to be prognostic in our study are objective, easy, and quick to measure, and do not require that the professional have special training or experience. The prediction of survival time may be improved by considering these variables. J Pain Symptom Manage 2006;31: Ó 2006 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Palliative care, prognostic factors, advanced cancer Address reprint requests to: Sofía Garrido Elustondo, MD, Gerencia de Atencion Primaria Area 7 Madrid, c/o Espronceda 24, 4 a planta, Madrid, Spain. Sgarrido.gapm07@salud.madrid.org Accepted for publication: October 24, Ó 2006 U.S. Cancer Pain Relief Committee Published by Elsevier Inc. All rights reserved. Introduction Doctor, how much longer will I live? Physicians who treat patients who suffer from incurable diseases may have been asked this question by the patient. More frequently, the patient s family may ask a similar question. The need to know this information is due /06/$--see front matter doi: /j.jpainsymman

2 486 de Miguel Sánchez et al. Vol. 31 No. 6 June 2006 to many spiritual and emotional factors: anticipation of grief, organizing the care, or saying goodbye. It may also be due to other more practical reasons, such as the wish to leave everything finished. An accurate estimation of survival allows the clinician to 1) adjust diagnostic and therapeutic measures to a realistic time frame to try to avoid undertreatment or overtreatment; 2) provide information to the patient and the family members that will allow them to manage their time and their emotional and economic resources; 3) organize the needed social resources; 4) plan the medical care as well as possible; 5) maximize the comfort of the patient; and 6) identify groups of patients with a similar prognosis for the purpose of designing clinical investigations. When establishing a prognosis, it is necessary to consider two main aspects: the nature of the disease and the patient s peculiarities. Many studies have analyzed variables that may predict survival time in cancer populations with advanced illness. These trials suggest that there is no association between the tumor s histology and the patient s survival time at this stage of the disease. 1 The parameters with a possible predictive value that have been studied up to now can be divided into five different groups: physician s impression, 2--6 functional status, clinical symptoms, indices to measure quality of life, and objective data. 9, Depending on the authors, contradictory results have been obtained. Recently, two prognostic indexes have been published: the Palliative Prognostic Index, described by Morita et al. 18 in 1999 in Japan, and the Palliative Prognostic Score, 19 which are models that use different variables that belong to these groups, with different systems of weighting. The utility of these scores in daily medical practice has not yet been defined The aim of the present study was to determine which clinical symptoms, which physical signs, and which characteristics that define the patient s functional status predict survival time in terminally ill cancer patients attended by one of Madrid s Home Care Support Teams. Methods Our investigation was a prospective longitudinal study that enrolled patients sent from specialized care or from primary care doctors and nurses to the Home Care Support Team of Madrid s Area 7. Patients had a diagnosis of terminally ill cancer, which according to the Spanish Society of Palliative Care means that life expectancy is less than 6 months. The study sample included all terminally ill cancer patients followed up by the Home Care Support Team of Madrid s Area 7 from January 1, 2001 until August 30, The patients were enrolled consecutively during this period of time. Madrid s Primary Care Area 7 is an urban area with an assigned population of approximately 500,000 inhabitants. When patients were enrolled, sociodemographic variables, variables related to the tumor, clinical symptoms and signs, and functional status were recorded. The sociodemographic variables were age and sex. The variables related to the tumor included localization, the existence of metastases, the number of metastases, the localization of the metastases, and the date of inclusion. The variables related to clinical symptoms and signs were anorexia (absence of appetite expressed by the patient); chronic nausea (unpleasant sensation that indicates the proximity of a vomit expressed by the patient); vomit (violent expulsion of the contents of the stomach through the mouth); state of the oral intake (normal or compromised); delirium (reference to the Diagnostic and Statistical Manual of Mental Disorders, 4th ed.); dyspnea at rest; hiccups; agitation (feeling of malaise with restlessness and increased activity with a certain degree of anxiety, fear, and tension observed by the caregiver); sleep disorders; illusions (hallucinations without objective object); hallucinations; level of consciousness (classified in four categories: alert [oriented in time, place, and person], apathetic [lethargic, forgetful, somnolent, passive, clumsy, lazy], confused [restless, aggressive, irritable, asleep], stuporous [disoriented in time, place, and person]); coherent language (correct connection of sentences during the interview); oriented in time (at every visit, aware of the year, month, and day); oriented in place (when asked, answers correctly); oriented in person (answers correctly with first and last name); heart rate; respiratory rate; temperature; and edema (presence or absence). The variables related to the functional status were the Palliative

3 Vol. 31 No. 6 June 2006 Survival Time in Terminally Ill Cancer Patients 487 Performance Status (PPS), 23 number of hours reclining per day, degree of dependence or independence in performing activities of daily living according to the Katz Index criteria, and level of activity (walks or is totally independent, walks with help, sits, bedridden). The PPS, which has 11 categories and uses values from 0 to 100, is a modification of the Karnofsky Performance Status (KPS) scale. The PPS takes into consideration five variables: walking around, activity and evidence of disease, independence for self-help, oral intake, and level of consciousness, 23 while the Karnofsky Scale only considers the first three variables. Patients were followed up with at least weekly visits until death. At each visit, the variables related to the clinical symptoms and signs, and to functional status, were recorded. The time of survival was defined as the number of days that passed between the initial visit and the date of death. Statistical Analysis The means, medians, and standard deviations (SD) were calculated for each of the quantitative variables, and the absolute and relative frequencies were calculated for the qualitative variables, with their corresponding 95% confidence intervals (CI). To evaluate which variables acted as survival prognostic factors in our population, a Cox univariate regression analysis was performed. Observation time for each patient was split into one or more periods, given the information available. Different time periods for the same patient were identified under the variable patient and this variable was used to allow for the lack of independence between observations coming from the same individual. The time of survival was defined as a dependent variable in every observation. The hazard ratio (HR), 95% CI, and statistical significance were calculated. P ¼ 0.05 was accepted as statistically significant. A multivariate analysis was then performed, incorporating in the model the variables that proved to be significant in the univariate analysis. The method used for model selection was backward stepwise elimination, in which the contrast for removal was based on the likelihood-ratio test. The STATA 7.0 program was used for these analyses. Table 1 Localization of Primary Tumor (n ¼ 98) Tumor Localization n % 95% CI Gastrointestinal Lung Genitourinary Head and neck Other Breast Unknown origin Liver, pancreas Central nervous system Total n ¼ absolute frequency; % ¼ relative frequency; CI ¼ confidence interval. Results Ninety-eight patients were studied. Sixty-one percent were men. The mean age was 72 years (SD years, range years). The localizations of the primary tumor are described in Table 1. Eighty-seven percent had metastases. The mean number of metastases was 1.5 (SD 1.2, range 0--4). Table 2 shows the metastases localization. The distribution of clinical symptoms and physical signs and distribution of the functional status at the first visit are shown in Tables 3 and 4, respectively. The typical patient with advanced cancer sent to the Home Care Support Team is between 60 and 84 years old and has normal oral intake. One of every five patients has dyspnea at rest, nausea, vomiting, delirium, edema, and/or insomnia. One-third of the patients have a PPS over 60, another third has a PPS of 50, and the remainder have a score of 40 or under. Nearly half of the patients spend more than 16 hours per day lying down and two-thirds have limited activity: they walk with help or with the Table 2 Localization of Metastases (n ¼ 98) Localization n % 95% CI Liver Lung Bone Lymph nodes Peritoneum Central nervous system Soft parts Pericardium Skin Others n ¼ absolute frequency; % ¼ relative frequency; CI ¼ confidence interval.

4 488 de Miguel Sánchez et al. Vol. 31 No. 6 June 2006 Table 3 Distribution of Clinical Symptoms and Physical Signs on the First Visit (n ¼ 98) Clinical Symptoms/Physical Signs n % 95% CI Anorexia Yes No Nausea Yes No Vomit Yes No State of oral intake Normal Compromised Delirium Yes No Dyspnea at rest Yes No Hiccups Yes No Agitation Yes No Sleep disorders Yes No Level of consciousness Alert Apathetic Confused Coma Coherent language Yes No Oriented in time Yes No Not able to evaluate Oriented in place Yes No Not able to evaluate Oriented in person Yes No Not able to evaluate Hallucinations Yes No Illusions Yes No Heart rate b.p.m $100 b.p.m Respiratory rate <24/minute $24/minute Temperature C C C $39 C Edema Yes n ¼ absolute frequency; % ¼ relative frequency; CI ¼ confidence interval.

5 Vol. 31 No. 6 June 2006 Survival Time in Terminally Ill Cancer Patients 489 Table 4 Distribution of the Characteristics That Define Functional Status on the First Visit (n ¼ 98) Functional Characteristic n % 95% CI Palliative Performance Status (PPS) Number of hours lying down per day < Bathing Independent Dependent Dressing Independent Dependent Toileting Independent Dependent Mobility Independent Dependent Incontinence Independent Dependent Nourishment Independent Dependent Level of activity Walks Walks with help Sits Limited to the bed n ¼ absolute frequency; % ¼ relative frequency; CI ¼ confidence interval. supervision of another person, or they can only remain lying or sitting. One of every five patients is independent for the basic activities of daily life, and another fifth of the patients are dependent for these activities. During the follow-up period, 250 evaluations were done in all, ranging between 1 and 11 entries per person. The median survival was 32 days (95% CI, ), with a minimum of one and a maximum of 213 days, resulting in a mean of days (95% CI, ). Table 5 shows the positive findings of the univariate analysis of the sociodemographic and clinical variables, and clinical symptoms and physical signs, Table 6 shows the positive findings of the characteristics that define the functional state of the patient. The following factors were associated with survival: anorexia; compromised oral intake; agitation; delirium; apathetic mental state; confused or in coma; coherent language; orientation in time, place, and person; hallucinations and/or illusions; heart rate; respiratory rate; PPS and Katz Index. When introducing the PPS in the multivariate analysis, the variables that are included in that index are excluded: number of hours lying down per day, level of activity, dependency for bathing, dressing, toileting, mobility, incontinence and eating, and compromise of the oral intake. In this multivariate analysis, three associated independent variables were identified: PPS of 50 or under, heart rate of 100 b.p.m. or more, and respiratory rate of 24/minute or more (Table 7). Discussion The selection of symptoms, physical signs, and functional characteristics permitted the development of a prognostic model that is objective, easy to measure, and independent of the professional s previous experience. It is

6 490 de Miguel Sánchez et al. Vol. 31 No. 6 June 2006 Table 5 Positive Findings of the Univariate Analysis of the Sociodemographic and Clinical Variables, and Clinical Symptoms and Physical Signs (n ¼ 250) Variable Median (Days) HR 95% CI P Anorexia No Yes Oral intake Normal 51 1 Compromised Delirium No 34 1 Yes Agitation No 34 1 Yes Level of consciousness Alert 49 1 Apathetic Confused Coma Coherent language Yes 34 1 No Oriented in time Yes 43 1 No Oriented in place Yes 39 1 No Oriented in person Yes 38 1 No Hallucinations and/or illusions No 38 1 Yes Heart rate b.p.m $100 b.p.m Respiratory rate <24/minute 39 1 $24/minute Temperature C C C $39 C Edema No 32 1 Yes HR ¼ hazard ratio; CI ¼ confidence interval; P ¼ grade of significance. important to point out that our study is based on patients living at home, while most publications refer to a hospitalized population. 18,24 However, at this stage of the disease, in our patients, the reasons for hospitalization are usually related to caregiver concerns, such as the lack of family support, the family s desire that the patient not die at home, or the absence of a caregiver. This fact allows the comparison with other publications. Regarding the prognostic variables, like other studies, survival time of studied population was independent of the age, sex, existence of metastases and their localization. 18,19 In our study, the variables with a significant predictive value were a PPS of 50 or less, a heart rate of 100 b.p.m. or more, and a respiratory rate of 24/minute or more. Other studies have not considered respiratory rate as a prognostic factor, and only Schonwetter et al. 25 state that the association of dependency for dressing oneself, dependency for mobility, anorexia, and heart rate are responsible for 30% of the survival variation. In spite of the widespread use in clinical practice of the KPS scale to evaluate performance status, we chose the PPS described by Anderson et al. 23 in Clinical experience with terminally ill cancer patients shows that oral intake and the level of consciousness, which are considered by the PPS but not by the KPS scale, are two important factors when measuring performance status. 23 The KPS scale has received criticism related to the absence of

7 Vol. 31 No. 6 June 2006 Survival Time in Terminally Ill Cancer Patients 491 Table 6 Univariate Analysis of the Characteristics That Define Functional State (n ¼ 250) Functional Characteristic Median (Days) HR 95% CI P Palliative $ Performance Status # Number of hours lying down per day < Bathing Independent 45 1 Dependent Dressing Independent 56 1 Dependent Toileting Independent 59 1 Dependent Incontinence Independent 43 1 Dependent Nourishment Independent 43 1 Dependent Level of activity Walks 71 1 Walks with help, sits, limited to bed HR ¼ hazard ratio; CI ¼ confidence interval; P ¼ grade of significance. operational parameters enabling the categorization of the work and self-help on a graduated scale, the unlimited inclusion of components that make up the scale, and the difficulty in knowing where to place up to 35% of the patients on the scale. 30,31 The PPS seems to detect fairly well the transition that the terminally ill cancer patient goes through before dying. The step from a PPS of 60 to 50 shows the change the patient suffers from being able to walk with help to being mainly sitting and/or lying. The step from 50 to 40 shows the change in which the patient is mainly in bed. 23 This fact is useful not only to predict the prognosis of these patients, but also to carry out studies, designate resources, and to know the need for medical care, nurse Table 7 Positive Findings of the Multivariate Analysis. Survival Analysis, Cox Regression (n ¼ 250) Variable HR 95% CI P PPS ¼ PPS # Heart rate $ 100 b.p.m Respiratory rate $ 24/ minute HR ¼ hazard ratio; CI ¼ confidence interval; P ¼ grade of significance. 2 log of the probability ¼ , P < care, nursing assistant s care, and/or help in the home. 23 Most studies use the KPS scale, however, despite the difficulties mentioned before, 19 and only Virik and Glare 32 make reference to the utility of the PPS as a prognostic indicator of survival in patients hospitalized in a palliative care unit in Sydney. The variables that were found to be prognostic in our study are objective, easy, and quick to measure, and do not require that the professional have special training or experience. The categories are clearly defined, do not require any invasive techniques or technology, and can be carried out by doctors and nurses. It would be convenient to compare these results with other populations of terminally ill cancer patients living at home. References 1. Reuben DB, Mor V, Hiris J. Clinical symptoms and length of survival in patients with terminal cancer. Arch Intern Med 1988;148: Oxeham D, Cornbleet MA. Accuracy of prediction of survival by different professional groups in a hospice. Palliat Med 1998;12: Vigano A, Dorgan M, Bruera E. The relative accuracy of prediction of clinical estimation of

8 492 de Miguel Sánchez et al. Vol. 31 No. 6 June 2006 duration of life for patients with end of life cancer. Cancer 1999;86(1): Christakis NA, Lamont EB. Extent and determinants of error in doctors prognoses in terminally ill patients: prospective cohort study. BMJ 2000;320: Glare P, Virik K, Jones M, et al. A systematic review of physician survival in terminally ill cancer patients. BMJ 2003;327: Garrido S, De Miguel C, Vicente F, et al. La impresión clínica como estimador de la supervivencia en pacientes oncológicos en situación terminal (Clinical impressions of terminal cancer patients as an estimator of time of survival). Aten Primaria 2004;34: Bruera E, Miller MJ, Kuehn N, Mac Eachern T, Hanson J. Estimate of survival of patients admitted to a palliative care unit: a prospective study. J Pain Symptom Manage 1992;7: Maltoni M, Pirovano M, Scarpi E, et al. Prediction of survival of patients all ill with cancer. Cancer 1995;75: Vigano A, Bruera E, Jhangri S, et al. Clinical survival predictors in patients with advanced cancer. Arch Intern Med 2000;160: Maltoni M, Pirovano M, Nanni O. Prognostic factors in terminal cancer patients. Eur J Palliat Care 1994;1: Addington-Hall JM, Mac Donald LD, Anderson HR. Can the Spitzer Quality of Index help to reduce prognostic uncertainty in terminal care? Br J Cancer 1990;62: Tamburini M, Brunelli C, Rosso S, et al. Prognostic value of quality of life scores in terminal cancer patients. J Pain Symptom Manage 1996;11: Morris JN, Suissa S, Sherwood S. Last days: a study of quality of life of terminal ill patients. J Chronic Dis 1986;39: Ralston SH, Gallacher SJ, Patel U. Cancer associated hypercalcemia: morbidity and mortality. Ann Intern Med 1990;112: Hermann FR, Safran CH, Levkoff SE, Minaker KL. Serum albumin level on admission as predictor of death, length of stay, and readmission. Arch Intern Med 1992;152: Maltoni M, Pirovano M, Nanni O. Biological indices predictive of survival in 519 Italian terminally ill cancer patients. J Pain Symptom Manage 1997;13: Alonso Martinez JL, Abinzano Guillen ML, Martinez Velasco ME. Morbimortalidad hospitalaria en la cuarta edad: identificación de factores pronóstico (Morbidity and mortality among the hospitalized aged. Identification of prognostic factors). Ann Med Int 1995;12: Morita T, Tsunoda J, Inoue S, Chihara S. The palliative prognostic index: a scoring system for survival prediction of terminally ill cancer patients. Support Care Cancer 1999;7: Pirovano M, Maltoni M, Nanni O, et al. A new palliative prognostic score: a first step for the staging of terminally ill cancer patients. Italian multicenter and study group on palliative care. J Pain Symptom Manage 1999;17: Maltoni M, Nanni O, Pirovano M, et al. Successful validation of the palliative prognostic score in terminally ill cancer patients. J Pain Symptom Manage 1999;17: Glare P, Virik K. Independent prospective validation of the PaP score in terminally ill patients referred to a hospital-based palliative medicine consultation service. J Pain Symptom Manage 2001;5: Morita T, Tsunoda J, Inoue S, Chihara S. Improved accuracy of physicians survival prediction for terminally ill cancer patients using the Palliative Prognostic Index. Palliat Med 2001;15: Anderson F, Downing M, Hill J, Casorso L, Lerch N. Palliative performance scale: a new tool. J Palliat Care 1996;12: Daas N. Estimating length of survival in end-- stage cancer: a review of the literature. J Pain Symptom Manage 1995;10: Schonwetter RS, Teasdate TA, Storey P. The terminal cancer syndrome. Arch Intern Med 1989;149: Yates JW, Chalmer B, McKeguey FP. Evaluation of patients with advanced cancer using the Karnofsky Performance Status. Cancer 1980;45: Mor V, Laliberte L, Morris LJ. The Karnofsky Performance Status Scale: an examination of its reliability and validity in a research setting. Cancer 1984;53: Verger E, Conill C, Chicote S, Azpiazu P. Valor del índice de Karnofsky como indicador del pronóstico de vida en pacientes oncológicos en fase terminal (Value of the Karnofsky Index as indicator of life prognosis in termainal oncology patients). Med Clin (Barc) 1993;100: Evans C, McCarthy M. Prognostic uncertainty in terminal care: can the Karnofsky Index help? Lancet 1985;1: Faris M. Clinical estimation of survival and impact of other prognostic factors on terminally ill cancer patients in Oman. Support Care Cancer 2003;11: Maltoni M, Nanni O, Derni S. Clinical prediction of survival is more accurate than the Karnofsky performance status in estimating life span of terminally ill cancer patients. Eur J Cancer 1994;30: Virik K, Glare P. Validation of the palliative performance scale for inpatients admitted to a palliative care unit in Sydney, Australia. J Pain Symptom Manage 2002;6: