Atropine, Hyoscine Butylbromide, or Scopolamine Are Equally Effective for the Treatment of Death Rattle in Terminal Care

Transcription

1 124 Journal of Pain and Symptom Management Vol. 38 No. 1 July 2009 Original Article Atropine, Hyoscine Butylbromide, or Scopolamine Are Equally Effective for the Treatment of Death Rattle in Terminal Care Hans Wildiers, MD, PhD, Chris Dhaenekint, Peter Demeulenaere, MD, Paul M.J. Clement, MD, PhD, Mark Desmet, MD, Rita Van Nuffelen, Jacques Gielen, MD, Erna Van Droogenbroeck, MD, Filip Geurs, MD, Jean-Pierre Lobelle, MD, and Johan Menten, MD, PhD, on behalf of the Flemish Federation of Palliative Care Department of General Medical Oncology (H.W., P.M.J.C., J.-P.L.), and Palliative Care Unit (P.M.J.C., R.V.N., J.M.), University Hospitals Leuven, Leuven; Palliative Care Unit (C.D., J.G.), A.Z. Sint-Elisabeth, Turnhout; University of Antwerp (P.D.), Antwerp; Palliative Care Unit (M.D.), A.Z. Virga Jesse, Hasselt; Palliative Care Unit (E.V.D.), Stedelijk Ziekenhuis, Aalst; and Palliative Care Unit (F.G.), St. Maria Hospital, Halle, Belgium Abstract Death rattle is a frequent symptom (25%e50%) in the terminal stage of life, but there is neither standardized treatment nor prospective investigation performed on the effectiveness of anticholinergic drugs. The aim of the present study was to investigate the effectiveness of three different anticholinergic drugs in the treatment of death rattle in the terminal stage of life. Terminal patients who developed death rattle were randomly assigned 0.5 mg atropine, 20 mg hyoscine butylbromide, or 0.25 mg scopolamine. Each treatment was initiated with a subcutaneous bolus, which was followed by continuous administration of the same drug. The intensity of death rattle and side effects were prospectively scored at different time points. Three hundred and thirty-three eligible patients were randomized to atropine, hyoscine butylbromide, or scopolamine after informed consent from the patient or the appointed representative. For the three drugs, death rattle decreased to a nondisturbing intensity or disappeared after one hour in 42%, 42%, and 37% of cases, respectively (P ¼ 0.72). Further, effectiveness improved over time without significant differences among the treatment groups (effectiveness at 24 hours was 76%, 60%, and 68%, respectively). In an analysis on the three groups together, treatment was more effective when started at a lower initial rattle intensity; median survival after start of therapy was 23.9 hours. These data suggest that Apart from a small unrestricted grant from Boehringer-Ingelheim of 500 euros, there was no financial support for this study, and no funding source had any role in study design, collection, analysis, interpretation of the data, or the writing of the report. H. Wildiers was the study chair, who conceived and designed the study and wrote the protocol in collaboration with J. Menten. C. Dhaenekint, P. Clement, M. Desmet, R. Van Nuffelen, J. Gielen, E. Van Droogenbroeck, and F. Geurs were active in the recruitment of eligible patients. P. Demeulenaere was Ó 2009 U.S. Cancer Pain Relief Committee Published by Elsevier Inc. All rights reserved. responsible for data management and site quality control. J.-P. Lobelle performed the statistical analysis. All of these authors assisted in the review of the report and approved the final manuscript. Address correspondence to: Hans Wildiers, MD, PhD, Department of General Medical Oncology/Multidisciplinary Breast Centre, University Hospitals Leuven, Herestraat 49, B-3000 Leuven, Belgium. hans.wildiers@uzleuven.be Accepted for publication: August 8, /09/$esee front matter doi: /j.jpainsymman

2 Vol. 38 No. 1 July 2009 Treatment of Death Rattle 125 there are no significant differences in effectiveness or survival time among atropine, hyoscine butylbromide, and scopolamine in the treatment of death rattle. J Pain Symptom Manage 2009;38:124e133. Ó 2009 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Death rattle, palliative care, terminally ill, scopolamine, atropine, hyoscine butylbromide, antimuscarinic drugs, anticholinergic therapy, hospice care Introduction Death rattle is a term used to describe the noise produced in dying patients by the oscillatory movements of secretions in the upper airways, in association with the inspiratory and expiratory phases of respiration. It is generally seen only in terminal patients with decreased consciousness or who are too weak to expectorate. Death rattle is a frequent clinical sign at the end of life, occurring in 25% 1 to 92% 2,3 of dying patients. Apart from nonpharmacological measures (semiprone position to encourage postural drainage, explanation of the pathophysiological processes to the relatives, gentle nasopharyngeal or tracheal suction), several antisecretory drugs have been used to treat this condition. These compounds block the production of bronchial secretions but do not affect existing respiratory secretions. Although death rattle is a frequent clinical sign in the terminal stage of life, there is no standardized treatment and few prospective investigations have been performed on the effectiveness of antisecretory drugs. Atropine is a widely available drug and is frequently used in home care for the treatment of death rattle. Scopolamine (hyoscine hydrobromide) is a muscarine receptor antagonist and is the epoxide of atropine; it has been reported to have a more potent effect on the inhibition of bronchial-secretion production and is less likely to cause tachycardia. 4 Scopolamine and atropine are tertiary amines, which pass the blood-brain barrier and can cause central effects, such as sedation, confusion, or paradoxical excitation, especially in the elderly. Hyoscine butylbromide is a semisynthetic derivative of scopolamine, which also is reported to be active in the treatment of respiratory tract secretions, 5 has peripheral effects similar to scopolamine, but is a quaternary ammonium derivative that does not pass through the blood-brain barrier. As a consequence, hyoscine butylbromide has no central adverse effects. There are a few reports on nonrandomized audits/studies comparing different anticholinergic drugs 6e8 and one small randomized study, 9 but results have been inconclusive. Studies generally have been too underpowered to draw relevant conclusions. The aim of this study was to assess, in a randomized multicenter prospective trial, the differences in effectiveness and side effects of three frequently used anticholinergic drugs (atropine, scopolamine, and hyoscine butylbromide) in the treatment of death rattle in terminal patients. Methods Patients and Procedures This was an open-label, prospective, randomized Phase III trial performed under the auspices of the Research Task Force of the Flemish Federation of Palliative Care. Randomization was done by a closed-envelope system and was stratified per center. Patients were recruited in six Flemish residential palliative care units (Belgium). Terminally ill patients aged 18 years or more were eligible for the study when death rattle of intensity score 1 or more developed. The rattle intensity score was defined as previously described: 0 ¼ not audible/1 ¼ only audible near the patient/ 2 ¼ clearly audible at the end of the patient s bed in a quiet room/3 ¼ clearly audible at a distance of about 9.5 m (at the door of the room) in a quiet room. 7 Patients were excluded if there were clear clinical indications of a secondary cause of rattle, called pseudodeath rattle, including respiratory infection, food/fluid aspiration, or cardiac failure with pulmonary edema. 10 In case of doubt about

3 126 Wildiers et al. Vol. 38 No. 1 July 2009 the cause of rattle, patients could be included, but signs of secondary causes were followed. Patients were randomly assigned to one of the three treatment possibilities. Group 1 received 0.5 mg atropine as a subcutaneous bolus, followed by 3 mg/24 hours (subcutaneous 0.5 mg every four hours, or continuous intravenous/subcutaneous infusion). This dose was based on the drug information leaflet and related publications. 11,12 Group 2 received 0.25 mg scopolamine as a subcutaneous bolus, followed by 1.5 mg/24 hours (subcutaneous 0.25 mg every four hours, or continuous intravenous/subcutaneous infusion). This dose was based on our previous experience 10 and is in the same range as suggested in 1997 by the British National Formulary. 13 Group 3 received 20 mg hyoscine butylbromide as a subcutaneous bolus, followed by 60 mg/24 hours (subcutaneous 10 mg every four hours, or continuous intravenous/subcutaneous infusion). 5 After treatment began, all patients were clinically evaluated by the nurse in charge at predefined time points: 30 minutes, one, four, 12, 24 hours, and then every 24 hours until death. If rattle persisted after 12 hours at a score of 2 or 3 in the absence of clear signs of pseudodeath rattle, it was advised to administer the starting bolus dose again, and thereafter to double the maintenance dose of the anticholinergic drug. If the death rattle persisted after 24 hours, further treatment was at the discretion of the treating physician. When obvious symptoms of pseudo-death rattle developed, the underlying pathology could be treated accordingly. In case of signs of possible intoxication (new development of acute glaucoma, flush, symptomatic tachycardia, central nervous system signs like psychosis, agitation, or delirium), further approaches were at the discretion of the investigator. Suction of respiratory secretions was allowed during the study. The primary objective was to observe a difference in effectiveness in the treatment of death rattle among the three anticholinergic drugs at one hour after the start of the treatment. The therapy was considered effective when the intensity of death rattle was lowered to intensity 0 or 1, because these scores are not considered clinically troublesome. With a persisting rattle intensity of 2 or 3, the treatment was considered ineffective. Patients with baseline rattle score of 1 were included, because it was expected that rattle often increases in time and that timely intervention might be beneficial to prevent increase. Secondary endpoints included differences in effectiveness at other time points, differences in side effects, time interval between start of treatment and death, and influence of underlying pathology on the effectiveness of the treatment. Ethical Approval The ethics committees of the University of Leuven (central committee) and the five participating centers approved the protocol. It was recommended to have informed consent from the patient, or if the patient was physically or mentally not capable of giving informed consent, to discuss informed consent with the dedicated representative who was appointed by the patient on entering the palliative care unit. At the first ethical approval (November 19, 2001), it was not specifically mentioned that, in case of force majeure (e.g., patient not capable of giving consent, and no legal representative present), patients could not be included. In a first phase of the study (2001e2003), 81 patients were included without informed consent. After international discussions with colleagues and further publications about this topic, 14 it was decided in 2003 to amend the protocol and to include only patients with informed consent (from patient or appointed representative), and to exclude the former randomized patients without informed consent from final analysis. The Ethics Committee of the University of Leuven approved this amendment in November 2006, before the analysis was initiated. Statistical Analysis Based on previous retrospective evidence, the response rate at one hour was estimated to be approximately 50%. 6,7 This study was powered to detect a 20% absolute difference in effectiveness after one hour of treatment, the primary endpoint among the three treatment arms, with an alpha value of 0.05 and power of Therefore, 91 patients per treatment arm were required. There was no central monitoring during the study, and thus, no view on the number of evaluable patients at one hour; therefore, it was decided to include a larger sample size (n ¼ 359 with informed consent; Fig. 1). The data were entered by one person (a data

4 Vol. 38 No. 1 July 2009 Treatment of Death Rattle 127 manager), and were later double-checked on site by one of the authors (P.D.). Statistical analysis was performed using the SAS version 8.2 software (SAS Institute Inc., Cary, NC). Descriptive parametric analysis and analysis of variance followed by a Tukey-Kramer test was used for continuous variables. Cross-tabulation followed by Chi-square test was used for discrete variables. Survival was described with the Kaplan-Meier method and tested with the log-rank test. Results Between November 2001 and November 2006, 440 patients were randomized in the study; 333 patients were evaluable. The most important reason for not being evaluable was the absence of informed consent (see earlier); rapid death was another reason for not being evaluable (Fig. 1). Informed consent was obtained in 68 cases from the patient, and in 265 cases from the appointed representative. Fig. 1. Trial profile. T þ 1 indicates time point one hour after the start of treatment. T þ 4 indicates time point four hours after start of treatment.

5 128 Wildiers et al. Vol. 38 No. 1 July 2009 Table 1 Patient and Tumor Characteristics at Baseline Characteristics Atropine (n ¼ 115) Hyoscine Butylbromide (n ¼ 106) Scopolamine (n ¼ 112) Total (n ¼ 333) Gender (n) Male Female Age, years Mean Median Diagnosis (n) Cancer Breast Central nervous system Gastrointestinal Gynecological Hematological Respiratory Urological Other Noncancer Tumor in the lung 66 (no) 63 (no) 66 (no) 195 (no) 49 (yes) 43 (yes) 46 (yes) 138 (yes) Tumor in central nervous system 82 (no) 65 (no) 91 (no) 238 (no) 33 (yes) 41 (yes) 21 (yes) 95 (yes) Initial rattle score at T 0 (n) Baseline patient and tumor characteristics were well balanced and are shown in Table 1. The mean age was over 70 years. Most of them were patients with cancer. Rattle score of 2 was the most frequent score at baseline. Effectiveness data after one hour were available for 315 of 333 patients. Treatment effectiveness at different time points is shown in Fig. 2. At the primary endpoint of one hour, there was no significant difference in effectiveness among atropine (42%), hyoscine butylbromide (42%), and scopolamine (37%) (Chi-square ¼ 0.66, degrees of freedom [DF] ¼ 2, P ¼ 0.72). It should be noted that, after one hour, more than half of the patients in each arm did not have a response. Differences among the three drugs at further time points also were not statistically different up to 48 hours. After 48 hours, it was not possible to make firm conclusions because of the small numbers of surviving patients. Figure 2 clearly shows that treatment effectiveness improves over time. There is a steady increase in effectiveness up to 24 hours, and a trend toward stabilization after 24 hours. At 24 hours, only about 30% of patients still had a rattle intensity of 2 or 3, whereas this number was about 70% at the start of anticholinergic therapy. However, as there was no placebo arm, it cannot be excluded that this would have been different without anticholinergic therapy. Effectiveness (defined as a rattle score of 0 or 1) was also evaluated post hoc as a function of the initial rattle score. Because there was no difference in global effectiveness among the three drugs, all data were taken together for this analysis. Figure 3 shows that when treatment is started at the low rattle score intensity of 1, treatment is much more effective than when started at a rattle score of 2 or 3 (Chisquare ¼ 72, DF ¼ 2, P < ), especially in the initial hours. For the small proportion of patients surviving more than 48 hours, this difference in effectiveness related to the initial intensity is no longer visible. For patients with a primary lung tumor or lung metastases, anticholinergic drug treatment was less effective (32%) compared with patients without tumor in the lung (46%, Chisquare ¼ 6.82, DF ¼ 1, P ¼ 0.009). There was no significant difference between patients with or without brain tumors or brain metastases.

6 Vol. 38 No. 1 July 2009 Treatment of Death Rattle PERCENTAGE EFFECTIVE Atropine Hyoscine butylbromide Scopolamine HOURS Fig. 2. Effectiveness per treatment arm over time. Therapy was considered effective when the intensity of rattle after treatment was 0 or 1, and was considered not effective if the score was 2 or 3. Numbers per bar indicate number of patients. Median survival was similar among the three treatment groups (Fig. 4). The median and mean survival values for the whole group were 23.9 and 39.2 hours, respectively, showing that death rattle is an objective sign of imminent death. Patients with higher baseline rattle scores die earlier: the median survival time was 38.6 hours for a baseline rattle score of 1, 20 hours for a score of 2, and 24.3 hours for a score of 3. There was no difference in effectiveness among different classes of primary tumors or among different age categories (data not shown). There were no significant differences in effectiveness at one hour among the different centers. A selection of the most important side effects at baseline, 12, and 24 hours is shown in Table 2. In general, no important PERCENTAGE EFFECTIVE baseline rattle score 1 baseline rattle score 2 baseline rattle score , HOURS Fig. 3. Overall effectiveness of anticholinergic treatment (all patients together) on rattle intensity at baseline and over time.

7 130 Wildiers et al. Vol. 38 No. 1 July 2009 Fig. 4. Overall survival after randomization per treatment arm. differences were present. Interestingly, consciousness decreased significantly more at 12 hours for scopolamine compared with the other two drugs (Chi-square ¼ 20.83, DF ¼ 8, P ¼ ), but this was not significant at 24 hours. Confusion measurements were only considered clinically relevant if consciousness was score 1 (alert) or 2 (somnolent). Only about one-quarter of patients were in this category. There were no important changes in confusion in these patients (numbers too small for statistics), but it is interesting that none were reported to have increased confusion at 12 and 24 hours. Clinical signs of cardiac failure, ileus, and bladder retention were rarely reported; at 12 hours, there was one case of clinical cardiac failure, one case of ileus, and six cases of bladder retention. Discussion This study is the first prospective randomized and multicenter study on the effectiveness of anticholinergic drugs in the treatment of death rattle in terminal palliative patients. There is no evidence for a significant difference in effectiveness after one hour or at other time points up to 48 hours among atropine, hyoscine butylbromide, and scopolamine. After two to three days, there was a nonsignificant advantage for atropine, but the numbers were too small to draw conclusions. Death rattle decreased to minor or absent intensity after one hour in about 40% of the patients, which confirms observations previously reported in retrospective studies, although the range and definition of effectiveness varied widely, and patient numbers were small in general. 12 Only a few studies have compared different anticholinergic drugs, none of which were prospectively randomized. In a Phase II study by Back et al., 7 scopolamine was more effective (56%) than glycopyrrolate (27%) in improving rattle score after 30 minutes (P ¼ 0.002) in 128 and 63 evaluable patients, respectively. However, after one hour and at the final score, this difference was not statistically different anymore. The starting dose of scopolamine of 0.4 mg subcutaneously was higher than that in our study (0.25 mg). A study by Hughes et al. 6 showed rather opposite results, because fewer patients responded to scopolamine (35%) compared with hyoscine butylbromide (54%) or glycopyrrolate (46%). Similar conclusions were reported in a recent smaller study by Hugel et al., 8 in which glycopyrrolate provided higher response rates in 36 patients matched with a group treated with scopolamine (100% vs. 78% response). The response rate in the latter study was very high compared

8 Vol. 38 No. 1 July 2009 Treatment of Death Rattle 131 Table 2 Safety Data: Differences in Side Effects Between Treatment Groups Item Atropine Hyoscine Butylbromide Scopolamine P -value Pulse (beats per minute; mean) T T þ 12 hours T þ 24 hours Temperature ( Celsius; mean) T T þ 12 hours T þ 24 hours Consciousness (n) Baseline at T 0 Alert ¼ Somnolent ¼ Stupor ¼ Coma ¼ Change T 0 / T þ 12 hours a Improved No change Worse Change T 0 / T þ 24 hours a Improved No change Worse Confusion adjusted for consciousness score of 1 or 2 b Baseline at T 0 No confusion ¼ Slight ¼ Severe ¼ Change T 0 / T þ 12 hours a Improved No change Worse d d d Change T 0 / T þ 24 hours a Improved No change Worse d d d T 0 indicates before treatment; T þ 12 hours indicates time point 12 hours after start of treatment; T þ 24 hours indicates time point 24 hours after start of treatment. a Change means that the score decreased or increased by at least 1 unit. b To be meaningful, the confusion scores were only analyzed when the consciousness scores were 1 or 2 at the time confusion was assessed. with other reports, but sample size was small and there was no randomization. It should be noted that the study by Hughes et al. 6 used a different definition of effectiveness, specifically, improvement, stabilization, or deterioration of rattle intensity. A rattle score going from intensity 3 to 2 would be a success according to this definition, but would not be called a clinically effective or relevant success. In our study, we decided to define success as rattle score being or becoming 0 or 1, because these scores are generally considered clinically tolerable. When our study was initiated, there was a supply problem with glycopyrrolate in Belgium, and future availability was not guaranteed at that moment. It was, therefore, decided to exclude glycopyrrolate and compare the three most frequently used drugs in Belgium for death rattledatropine (widespread availability), scopolamine (often used in hospital settings), and hyoscine butylbromide (with the potential advantage of not causing central effects). Our study shows that effectiveness of the three drugs is similar at the presently used doses. In general, side effects were similar (Table 2), although scopolamine decreased consciousness more than hyoscine butylbromide (as expected) and atropine (which is not really expected). Decreased consciousness can be desired or not depending on the clinical situation, and might play a role in the choice of the anticholinergic drug. A question that arises is whether this study should not have included a placebo arm. It is theoretically possible that rattle decreases

9 132 Wildiers et al. Vol. 38 No. 1 July 2009 spontaneously when death approaches. Likar found no difference in effectiveness between placebo and scopolamine, but the number of patients included was small and there was no informed consent in this small study. 9 It is ethically debatable to include a placebo arm for symptom control in dying patients because many observational reports have suggested effectiveness of anticholinergic therapy, and currently, these drugs are generally accepted as standard treatment of death rattle in the palliative care community. Moreover, our study clearly shows (Fig. 2) that effectiveness increases in time, mainly for initial intensity scores of 2 and 3, and that it takes longer for rattle with a high intensity score to decrease compared with lower scores. Thus, although the lack of a placebo arm precludes direct evidence of the efficacy of anticholinergic therapy, the previous arguments suggest effectiveness. It is questionable if it will ever be ethically acceptable and/or feasible to perform a placebo-controlled trial in dying patients. A possible limitation of this study is that it was not blinded and that the same nurse administered the medication and scored the rattle. Unfortunately, there was no other practical solution when the study was designed. There was also no systematic recording of intravenous and oral fluid intake, which might have influenced the development and evolution of the rattle. However, there is, in general, a very restrictive use of extra fluid administration in palliative care units in Belgium, and only a few patients probably received significant amounts of intravenous or subcutaneous fluid during their stay in the palliative care unit. Several conclusions are relevant for clinical practice. There is no evidence for a significant difference among atropine, hyoscine butylbromide, and scopolamine at the presently used dosage. No important changes in side effects have been observed, although consciousness decreased more with scopolamine compared with hyoscine butylbromide and atropine. It can be recommended to use the antimuscarinic drug the treating physician is most familiar with, but the desired effect on consciousness also may play a role in treatment choice. This therapy can easily be administered at home or in nursing homes. Further, antimuscarinic treatment is more effective when started early. We recommend starting treatment at the first occurrence of rattle, even at low intensity, because effectiveness is higher for rattle with lower intensity. We also recommend continuing treatment because effectiveness improves in time. Death rattle also is a strong predictor of impending death, because the median survival in our trial was only 23 hours. Life expectancy is important for relatives in making decision to stay at the bedside of their family member during his or her final hours. Finally, this study shows that it is feasible to perform prospective randomized studies in dying patients. Acknowledgments The authors thank all the study centers and patients for their voluntary cooperation, Bart Van den Eynden (Chair of Palliative Care, University of Antwerp) for funding the data management, and Ulla van Leeuw for data management. References 1. Lichter I. The last 48 hours of life. J Palliat Care 1990;6:7e Ellershaw JE, Sutcliffe JM, Saunders CM. Dehydration and the dying patient. J Pain Symptom Manage 1995;10(3):192e Morita T, Tsunoda J, Inoue S, Chihara S. Risk factors for death rattle in terminally ill cancer patients: a prospective exploratory study. Palliat Med 2000;14(1):19e O Donnell V. Symptom management. The pharmacological management of respiratory tract secretions. Int J Palliat Nurs 1998;4(4):199e Bausewein C. Comparative costs of hyoscine injections. Palliat Med 1995;9: Hughes A, Wilcock A, Corcoran R, Lucas V, King A. Audit of three antimuscarinic drugs for managing retained secretions. Palliat Med 2000; 14(3):221e Back IN, Jenkins K, Blower A, Beckhelling J. A study comparing hyoscine hydrobromide and glycopyrrolate in the treatment of death rattle. Palliat Med 2001;15(4):329e Hugel H, Ellershaw J, Gambles M. Respiratory tract secretions in the dying patient: a comparison between glycopyrronium and hyoscine hydrobromide. J Palliat Med 2006;9(2):279e Likar R. Klinische Untersuchung ûber die Wirkung von Scopolamin-Hydrobromicum beim terminalen Rasseln (randomisierte, doppelblind, plazebokontrollierte studie)[study about the

10 Vol. 38 No. 1 July 2009 Treatment of Death Rattle 133 efficacy of scopolamin-hydrobromicum in terminal rattle (randomized, double-blind, placebo-controlled study)]. Z Palliativmed 2002;3:15e Wildiers H, Menten J. Death rattle: prevalence, prevention and treatment. J Pain Symptom Manage 2002;23(4):310e Cowl CT, Prakash UBS, Kruger BR. The role of anticholinergics in bronchoscopyda randomized clinical trial. Chest 2000;118:188e Mirakhur RK, Dundee JW. Comparison of the effects of atropine and glycopyrrolate on various end-organs. J R Soc Med 1980;73:727e British National Formulary. Prescribing in palliative care, 33rd ed. London: RPS Publishing, Karlawish JHT. Research involving cognitively impaired adults. N Engl J Med 2003;348(14): 1389e1392.