Effects of Parenteral Hydration in Terminally Ill Cancer Patients: A Preliminary Study
http://www.100md.com
《临床肿瘤学》
The University of Texas M.D. Anderson Cancer Center, Houston, TX
Hospital Eva Peron, Rosario, Argentina
Instituto Nacional de Cancer, Santiago, Chile
Clinica de Dolor, Fundacion Santafe de Bogotá, Bogotá, Columbia
Centro Regional de Medicina, Los Montalvos, Spain
ABSTRACT
PURPOSE: Most patients with cancer develop decreased oral intake and dehydration before death. This study aimed to determine the effect of parenteral hydration on overall symptom control in terminally ill cancer patients with dehydration.
PATIENTS AND METHODS: Patients with clinical evidence of mild to moderate dehydration and a liquid oral intake less than 1,000 mL/day were randomly assigned to receive either parenteral hydration with 1,000 mL (treatment group) or placebo with 100 mL normal saline administered over 4 hours for 2 days. Patients were evaluated for target symptoms (hallucinations, myoclonus, fatigue, and sedation), global well-being, and overall benefit.
RESULTS: Twenty-seven patients randomly assigned to the treatment group had improvement in 53 (73%) of their 73 target symptoms versus 33 (49%) of 67 target symptoms in the placebo group (n=22; P = .005). Fifteen (83%) of 18 and 15 (83%) of 18 patients had improved myoclonus and sedation after hydration versus eight (47%) of 17 and five (33%) of 15 patients after placebo (P = .035 and P = .005, respectively). There were no significant differences of improvement in hallucinations or fatigue between groups. When blinded to treatment, patients (17 [63%] of 77) and investigators (20 [74%] of 27) perceived hydration as effective compared with placebo in nine (41%) of 22 patients (P = .78) and 12 (54%) of 22 investigators (P = .15), respectively. The intensity of pain and swelling at the injection site were not significantly different between groups.
CONCLUSION: Parenteral hydration decreased symptoms of dehydration in terminally ill cancer patients who had decreased fluid intake. Hydration was well tolerated, and a placebo effect was observed. Studies with larger samples and a longer follow-up period are justified.
INTRODUCTION
Decreased oral intake, a frequent complication of advanced cancer, results from various causes, including profound anorexia, odynophagia, oral cavity lesions, dysphagia, nausea and vomiting, delayed gastric emptying, bowel obstruction, cognitive impairment, and severe mood disorders.1-4 Patients with advanced cancer who have dehydration or decreased oral intake almost always receive parenteral hydration in acute care facilities but almost never in hospices.3-5 In the past decade, the consequences of dehydration in terminally ill cancer patients have generated strong debate, with arguments for and against fluid administration.5-7
The decision of whether to administer fluids should be individualized, based on a careful assessment of a patient's clinical presentation, the potential advantages of parenteral fluids, and the patient's and family's wishes.6-8 Unfortunately, the decision-making process is complicated by the absence of randomized controlled trials focusing on the potential advantages and disadvantages of parenteral hydration.9 However, retrospective studies suggest that hydration can reduce neuropsychiatric symptoms such as sedation, hallucinations, myoclonus, and agitation.10,11
Methods of fluid hydration for palliative care patients include subcutaneous hydration or hypodermoclysis.6,12-14 This method has many potential advantages over other methods of fluid replacement: it may be started and stopped with no risk of thrombosis or bleeding; it is easier to manage in the home setting, with administration performed by family members or the patients themselves after minimal training; infusions can be easily administered by gravity, thereby avoiding the need for infusion pumps; and the same infusion site can be used for many days.6,13
The purpose of this randomized, controlled, double-blind study was to determine the effects of parenteral hydration with 1,000 mL/d versus 100 mL/d of normal saline, administered intravenously or subcutaneously, on overall symptom control in patients with advanced cancer.
PATIENTS AND METHODS
Study Design
This study was conducted at The University of Texas M.D. Anderson Cancer Center, Houston, TX; Hospital Eva Peron, Rosario, Argentina; Instituto Nacional de Cancer, Santiago, Chile; Clinica de Dolor, Fundacion Santafe de Bogotá, Bogotá, Columbia; and Centro Regional de Medicina, Los Montalvos, Spain. Patients had to meet the following inclusion criteria to be admitted to the study: a diagnosis of advanced cancer, defined as locally recurrent or metastatic, with no further treatment planned; an oral intake of less than 1,000 mL/d, as determined by clinical assessment; and evidence of mild to moderate dehydration, exhibited by decreased turgor in the subclavicular region lasting more than 2 seconds. In addition, patients had to have one or more of the following findings: dry mouth; thirst; decreased volume of urine output, as reported by the patient; a darker color of urine than usual, in the absence of reasons for jaundice or hematuria; and laboratory values consistent with dehydration, such as an elevated blood urea nitrogen to creatinine ratio of more than 20:1, when this value was obtained within 24 hours of admission to the study. Finally, patients had to be older than 16 years, able to understand and give consent for participation in the study, and able to tolerate parenteral treatment and the application of a subcutaneous or intravenous cannula. Exclusion criteria included a patient's refusal to participate; the presence of severe dehydration, defined as a decreased systolic resting blood pressure of 30 mmHg or lower from the patient's baseline value; low perfusion of the limbs; no urine output for 12 hours or longer; a decreased level of consciousness; or evidence of severe renal failure or bilateral hydronephrosis.
The study was carried out after institutional review boards at each study site approved the study and patients gave written informed consent to participate. All participating investigators underwent a 2-day training session at M.D. Anderson Cancer Center, during which the protocol was reviewed, including the method of administering hydration and the appropriate blinding of the type of parenteral infusion given. Random numbers determining treatment allocation were calculated at M.D. Anderson Cancer Center and delivered to the investigators in sealed envelopes. The randomization code was kept confidential at M.D. Anderson Cancer Center until the completion of the study.
The study was randomized, controlled, and double blind. Patients who met the inclusion criteria and agreed to participate were randomly assigned to receive either 1,000 mL of normal saline (treatment group) or 100 mL of normal saline (placebo group) as an infusion over 4 hours for 2 days. Both types of infusion were given parenterally (either intravenously or subcutaneously). Patients who already had an intravenous access device (n = 12) received infusions intravenously, and patients with no intravenous access device received infusions subcutaneously (n = 37). One investigator at each institution was unblinded and was responsible for preparing the saline, the droppers, and the portable and nonportable pumps to maintain the blinding. The investigator responsible for assessing the patient and the patient himself or herself were not aware of the amount of fluid being administered. In all cases, at least one test of blinding was conducted before activation of the study at each institution.
Outcome Measures
No previously defined outcomes for our study existed because no previous randomized controlled trials of hydration had occurred. We chose four target symptoms for hydration (sedation, fatigue, hallucinations, and myoclonus) on the basis of retrospective studies10-12,14,15 and our clinical experience. These target symptoms were found to be appropriate at a workshop at M.D. Anderson that included the study investigators and six other nonparticipating palliative care researchers from Houston. Patients scored the presence of the four target symptoms on a numeric scale of 0 (absent) to 10 (worst possible). The symptoms were considered present whenever the score was 1 or greater, and improvement was defined as a decrease of one point or more. The final scores were then added for the treatment (1,000 mL/d) and the placebo (100 mL/d) groups.
The main outcome of the study for the purpose of sample size calculation was the global assessment of the overall benefit of hydration to the patient, as determined by the physician and patient on day 2. The study was designed to detect a difference in the proportion of patients in the two groups having any important overall benefit. Parenteral hydration with 1,000 mL of normal saline was considered successful if 75% of the patients were perceived as having any important benefit (ie, a score of 3 or more) at the end of 2 days of hydration. We expected that approximately 50% of the patients receiving only 100 mL of normal saline would also be perceived as having an important benefit.
Our second proposed method of analysis was to test the two groups separately to determine whether the proportion of patients perceived to have some benefit was equal to 50% or greater than 50%.
The planned sample size was calculated as 54 patients per group to allow for the detection of differences, with an 80% power and a one-sided significance level of .05.
Assessments
The following six assessments were performed before parenteral hydration began (baseline) and at the end of days 1 and 2, unless otherwise noted: (1) a history, physical examination, and review of medications at baseline; (2) a Mini-Mental State Examination for assessment of cognitive function16 at baseline and on day 2; (3) an evaluation for target symptoms of dehydration, including hallucinations, myoclonus, fatigue, and sedation, during the previous 24 hours was assessed by an investigator who was unaware of the type of treatment that the patient received; these four symptoms were assessed using a numeric scale of 0 (no hallucination, no fatigue, and so on) to 10 (worst possible hallucinations, worst possible fatigue, and so on); (4) an assessment of the patient's well-being performed by both the patient and investigator at baseline and on day 2 using a numeric scale of 0 (worst possible) to 10 (best possible); (5) a global assessment of overall benefit on day 2 performed by the patient and investigator using a numeric scale of 1 (no important benefit) to 7 (greatly important benefit); (6) an assessment for toxicity (eg, pain, swelling, or leakage at the site of infusion) in each patient who received subcutaneous hydration using a numeric scale of 0 (no toxicity) to 10 (worst toxicity).
After the assessments on day 2, the physician and patient were free to choose to continue their preferred method of hydration. The brief 2-day period of observation was established to allow the treating physicians to provide their preferred method of hydration as rapidly as possible, after a sufficient time had passed for clinically detectable changes to occur. If, in the opinion of the treating physician, a patient experienced acute deterioration during the 2-day period, the patient was removed from the study and provided standard care.
RESULTS
The planned sample size of this study was 108 patients. However, because of difficulties in identifying patients with dehydration who had normal cognition and were willing to participate in the study, the total accrual was only 51 patients at the time of study closure. Twenty-eight patients were randomly assigned to the treatment group and 23 patients were randomly assigned to the placebo group (Fig 1). Only 49 (96%) of these patients were assessable; one patient from each group did not complete the study because of family refusal in one case and dyspnea and rapid deterioration in the other.
Table 1 lists the characteristics of the patients admitted to the study. All patients were receiving opioid analgesic agents for pain management.
Table 2 lists the effects of hydration on the target symptoms. Fifty-three (73%) of 73 target symptoms experienced by the treatment group improved, compared with 33 (49%) of 67 target symptoms in the placebo group (P = .006).
Patients blindly perceived hydration as effective in 17 (63%) of 27 cases in the treatment group and in nine (41%) of 22 cases (P = .78) in the placebo group. Investigators, who were unaware of the type of hydration patients received, perceived hydration as effective in 20 (74%) of 27 cases in the treatment group and 12 (54%) of 22 cases (P = .15) in the placebo group.
According to our study design, if 75% of patients had any perceived important benefit at the end of 2 days of hydration with 1,000 mL of normal saline, the treatment would be considered successful. We achieved this goal, with 21 (78%) of 27 patients in the treatment group having a perceived important benefit of the treatment. However, 13 (59%) of 22 patients in the placebo group also had a perceived important benefit of the treatment, which was higher than the expected 50%. The difference in the proportion of patients in the two groups who had a perceived important benefit was not statistically significant (P = .16). We had a 42% power to declare these proportions significantly different with our reduced numbers of 27 and 22 patients per group. Notably, if we had enrolled the original 54 patients per group and had found the same proportions (78% v 59%), we also would have had a reduced power of 61% to declare the difference in proportions statistically significant.
Our second proposed method of analysis was to test the two groups separately to determine whether the proportion of patients perceived to have some benefit was equal to 50% or greater than 50%. The results in the treatment group caused us to reject the null hypothesis (P = .0035), whereas those in the placebo group did not (P = .20). In other words, the proportion of patients in the treatment group judged by the treating physicians to have received a benefit was greater than 50%, with a significance level of .0035. In contrast, the proportion of patients in the placebo group judged by the treating physicians to have received a benefit was not significantly different from 50%.
The mean score (on a scale of 1 to 7) given by the investigators for their overall perception of the importance of the treatment benefit was 4.5 ± 2.3 in the treatment group and 3.4 ± 2.4 in the placebo group (P = .13; Table 3). The mean score given by patients was 3.8 ± 2.2 in the treatment group and 3.6 ± 2 in the placebo group (P > .2). The mean scores for the perception of an overall sensation of well-being are also provided in Table 3.
With subcutaneous administration of saline, the mean intensities of pain at the injection site were 2.10 ± 2.95 and 1.75 ± 2.55 for the treatment (n = 20) and placebo17 group (P = .1), respectively; the mean scores for injection-site swelling were 0.82 ± 1.13 and 1.41 ± 1.66 (P = .64). There were no significant differences in Mini Mental State Examination score at baseline and day 2 between treatment and placebo group.
DISCUSSION
In this randomized, controlled, double-blind study, we compared the effects of hydration with either 1,000 or 100 mL of normal saline on target symptoms and overall perception of treatment effectiveness by patients and investigators after 2 days of treatment. To our knowledge, this is the first randomized controlled trial comparing hydration with placebo in patients with advanced cancer. Our results suggested that hydration was able to improve the combined target symptom score and decrease myoclonus and sedation in the treatment group compared with the placebo group. A trend toward a perception of overall benefit for patients in the treatment group was also observed. Unfortunately, these results are far from conclusive and need to be confirmed by other studies.
Patient enrollment in this study was much more difficult than we had anticipated. At the time of their clinical dehydration and/or decreased oral intake, patients had frequently already developed delirium and were therefore unable to give consent for participation in the study. Future studies should attempt to obtain consent for participation before this complication occurs or use a design that allows proxy consent in patients who have already developed delirium.
Our findings suggested that double-blind studies are feasible and that patients identified for such studies are generally willing to give consent. One of the major challenges for such studies will be to obtain appropriate funding for conducting these studies, which are unlikely to be of interest to the pharmaceutical industry and so will have to compete for the limited funding allocated by granting agencies for clinical trials in palliative care.17
Some investigators were concerned about the possibility of clinical deterioration in the placebo group in a study conducted over multiple days. Patients, without receiving hydration, are likely to develop neurologic, renal, and cardiovascular complications as a result of the rapid decrease in intravascular volume. Clinicians who routinely administer parenteral hydration and believe there are clinical benefits from hydration have ethical concerns about randomly assigning patients to a placebo group. Future clinical trials of parenteral hydration should ideally be conducted in settings where hydration is not routinely administered, such as hospices. Therefore, the final assessment was conducted at the end of the infusion on day 2, which meant that the final assessment compared the differences between the treatment and placebo groups over a period of less than 36 hours. It is likely that much more significant differences will occur over time. Because the outcomes of this study were mostly of a neuropsychiatric nature, future studies should include more investigator assessments, given that delirium is an extremely frequent event in patients with advanced cancer and dehydration and that such patients may not be able to complete their own assessments.18 Family assessments regarding neuropsychiatric symptoms may be an important outcome as patients become progressively unresponsive. The lack of changes in cognition as measured by the Mini Mental State Examination is not surprising because normal cognition was a criterion for eligibility in these patients, and therefore there was very limited room for improvement. The outcomes chosen for this study were those found to be influenced by hydration in previous retrospective studies.10-12,14 Dry mouth and thirst were not included because previous research has shown that the association between these symptoms and the hydration status of patients with cancer is limited.19,20
The fact that 59% of patients in the placebo group perceived important overall symptomatic benefit after less than 36 hours emphasizes the need for double-blind studies. Our group has observed a large placebo effect in other symptom control studies in advanced cancer patients.21-24 It is possible that frequent contact with investigators results in improved symptom expression. Findings of this study and our previous experience21-24 suggest that a placebo control is very important even for preliminary symptom control studies. Our findings suggested that this blind perception of overall effectiveness might not be sensitive enough to detect the effects of full hydration compared with specific symptom assessments. The most effective outcomes will need to be carefully defined in future clinical trials. Prospective pilot studies of rehydration conducted on an open basis may help better delineate target symptoms for future randomized clinical trials.
Dehydration can cause various severe symptoms that can be difficult to interpret in terminally ill cancer patients because similar symptoms can be caused by the cancer itself or by the drugs the patient is taking25; such symptoms include dry mouth, cognitive failure, myoclonus, hallucinations, hyperalgesia, and grand mal seizures.26,10 Dehydration can also result in fatigue, nausea, postural hypotension, fever with no underlying infectious process, increased risk for bedsores, and constipation.26-28 On the other hand, parenteral hydration can increase the risk for edema and respiratory distress and requires that patients remain attached to an infusion device and undergo venipunctures. In addition, intravenous hydration is costly and difficult to maintain at home and, therefore, frequently requires patients to remain in the hospital until death.4-6
Our study found that hydration resulted in rapid decreases in the level of sedation, and myoclonus and a trend toward a decreased level of hallucinations. These benefits may have resulted from the hydration per se or simply from an increased elimination of active opioid metabolites from our patients, all of who were receiving opioids for their pain.15,18,25 This issue can be addressed in future studies by stratifying patients according the use of opioid therapy. From a pragmatic perspective, however, given that more than 80% of terminally ill cancer patients receive opioids, future clinical trials should focus mostly on patients such as those in our study.
Our findings suggest that clinical symptom improvement can be observed with a volume of hydration of approximately 1,000 mL/d. Previous research has shown that patients treated by palliative care teams receive significantly smaller volumes of hydration compared with those admitted to cancer centers.29 Subcutaneous hydration with a total daily volume of approximately 1,000 mL can be easily achieved at home, can be administered by patients and family members independently,6,9,14 and—as our preliminary findings suggest—can result in measurable symptomatic improvement.
Few areas of clinical care involve such disparity in clinical practice, such passionate rhetoric, and such lack of evidence as hydration near the end of life.5-9,15,20,26 Appropriately powered, randomized, double-blind studies with longer follow-up periods and more investigator-measured clinical outcomes than we used are required to help resolve some of these issues.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported by the Brown Foundation, Houston, TX; and the Tobacco Settlement Foundation.
Presented in part at the 40th Annual Meeting of the American Society of Clinical Oncology, New Orleans, LA, June 5-8, 2004.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Ripamonti C, Gemlo BT: Methods for artificial feeding and drainage of the gastrointestinal tract, in Bruera E, Higginson I (eds): Cachexia-Anorexia in Cancer Patients. New York, NY, Oxford University Press, 1996, pp 141-157
Alexander HR, Norton JA: Pathophysiology of cancer cachexia, in Doyle D, Hanks G, MacDonald N (eds): Oxford Textbook of Palliative Medicine. New York, NY, Oxford University Press, 1993, pp 316-329
Bruera E, Miller L, McCallion J, et al: Cognitive failure in patients with terminal cancer: A prospective study. J Pain Symptom Manage 7:192-195, 1992
Coyle N, Adelhardt J, Foley KM, et al: Character of terminal illness in the advanced cancer patient: Pain and other symptoms during the last four weeks of life. J Pain Symptom Manage 5:83-93, 1990
Billings JA: Comfort measures for the terminally ill: Is dehydration painful J Am Geriatr Soc 33:808-810, 1985
Fainsinger RL, MacEachern T, Miller MJ, et al: The use of hypodermoclysis for re-hydration in terminally ill cancer patients. J Pain Symptom Manage 9:298-302, 1994
Dunphy K, Finlay L, Rathbone G, et al: Rehydration in palliative and terminal care: If not, why not Palliat Med 9:221-228, 1995
Dunlop RJ, Ellershaw JE, Baines M, et al: On withholding nutrition and hydration in the terminally ill: Has palliative medicine gone too far A reply. J Med Ethics 21:141-143, 1995
Bruera E, MacDonald N: To hydrate or not to hydrate: How should it be J Clin Oncol 18:1156-1158, 2000
Bruera E, Franco JJ, Maltoni M, et al: Changing pattern of agitated impaired mental status in patients with advanced cancer: Association with cognitive monitoring, hydration and opioid rotation. J Pain Symptom Manage 10:287-291, 1995
de Stoutz ND, Bruera E, Suarez-Almazor M: Opioid rotation (OR) for toxicity reduction in terminal cancer patients. J Pain Symptom Manage 10:378-384, 1995
Bruera E, Brenneis C, Michaud M, et al: Use of the subcutaneous route for the administration of narcotics in patients with cancer pain. Cancer 62:407-411, 1988
MacMillan K, Bruera E, Kuehn N, et al: A prospective comparison study between a butterfly needle and a Teflon cannula for subcutaneous narcotic administration. J Pain Symptom Manage 9:82-84, 1994
Fainsinger RL, Bruera E: Hypodermoclysis for symptom control versus the Edmonton Injector. J Palliat Care 7:5-8, 1991
Lawlor PG: Delirium and dehydration: Some fluid for thought Support Care Cancer 10:445-454, 2002
Folstein MF, Folstein SE, McHugh P: "Mini-Mental State": A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12:189-198, 1975
Field MJ, Cassel CK: Approaching Death: Improving Care at the End of Life. Washington, DC, National Academy Press, 1997
Lawlor PG, Gagnon B, Mancini IL, et al: Occurrence, causes, and outcome of delirium in patients with advanced cancer: A prospective study. Arch Intern Med 160:786-794, 2000
Lawlor PG, Nekolaichuk C, Gagnon B, et al: Clinical utility, factor analysis, and further validation of the memorial delirium assessment scale in patients with advanced cancer: Assessing delirium in advanced cancer. Cancer 88:2859-2867, 2000
Morita T, Tei Y, Tsunoda J, et al: Determinants of the sensation of thirst in terminally ill cancer patients. Support Care Cancer 9:177-186, 2001
Bruera E, Roca E, Cedaro L, et al: Action of oral methylprednisolone in terminal cancer patients: A prospective randomized double-blind study. Cancer Treat Rep 69:751-754, 1985
Eduardo E, Moyano J, Sala R, et al: Dexamethasone in addition to metoclopramide for chronic nausea in patients with advanced cancer: A randomized controlled trial. J Pain Symptom Manage 28:381-388, 2004
Bruera E, Strasser F, Palmer JL, et al: Effect of fish oil on appetite and other symptoms in patients with advanced cancer and anorexia/ cachexia: A double-blind, placebo-controlled study. J Clin Oncol 21:129-134, 2003
Bruera E, Chadwick S, Brenneis C, et al: Methylphenidate associated with narcotics for the treatment of cancer pain. Cancer Treat Rep 71:67-70, 1987
Steiner N, Bruera E: Methods of hydration in palliative care patients. J Palliat Care 14:6-13, 1998
Burge FI: Dehydration symptoms of palliative care cancer patients. J Pain Symptom Manage 8:454-464, 1993
de Conno F, Ventafridda V, Saita L: Skin problems in advanced and terminal cancer patients. J Pain Symptom Manage 6:247-256, 1991
Sykes N: Constipation and diarrhea, in Doyle D, Hanks G, MacDonald N (eds): Oxford Textbook of Palliative Medicine. New York, NY, Oxford University Press, 1993, pp 299-310
Bruera E, Belzile M, Watanabe S, et al: Volume of hydration in terminal cancer patients. Support Care Cancer 4:147-150, 1996(Eduardo Bruera, Raul Sala)
Hospital Eva Peron, Rosario, Argentina
Instituto Nacional de Cancer, Santiago, Chile
Clinica de Dolor, Fundacion Santafe de Bogotá, Bogotá, Columbia
Centro Regional de Medicina, Los Montalvos, Spain
ABSTRACT
PURPOSE: Most patients with cancer develop decreased oral intake and dehydration before death. This study aimed to determine the effect of parenteral hydration on overall symptom control in terminally ill cancer patients with dehydration.
PATIENTS AND METHODS: Patients with clinical evidence of mild to moderate dehydration and a liquid oral intake less than 1,000 mL/day were randomly assigned to receive either parenteral hydration with 1,000 mL (treatment group) or placebo with 100 mL normal saline administered over 4 hours for 2 days. Patients were evaluated for target symptoms (hallucinations, myoclonus, fatigue, and sedation), global well-being, and overall benefit.
RESULTS: Twenty-seven patients randomly assigned to the treatment group had improvement in 53 (73%) of their 73 target symptoms versus 33 (49%) of 67 target symptoms in the placebo group (n=22; P = .005). Fifteen (83%) of 18 and 15 (83%) of 18 patients had improved myoclonus and sedation after hydration versus eight (47%) of 17 and five (33%) of 15 patients after placebo (P = .035 and P = .005, respectively). There were no significant differences of improvement in hallucinations or fatigue between groups. When blinded to treatment, patients (17 [63%] of 77) and investigators (20 [74%] of 27) perceived hydration as effective compared with placebo in nine (41%) of 22 patients (P = .78) and 12 (54%) of 22 investigators (P = .15), respectively. The intensity of pain and swelling at the injection site were not significantly different between groups.
CONCLUSION: Parenteral hydration decreased symptoms of dehydration in terminally ill cancer patients who had decreased fluid intake. Hydration was well tolerated, and a placebo effect was observed. Studies with larger samples and a longer follow-up period are justified.
INTRODUCTION
Decreased oral intake, a frequent complication of advanced cancer, results from various causes, including profound anorexia, odynophagia, oral cavity lesions, dysphagia, nausea and vomiting, delayed gastric emptying, bowel obstruction, cognitive impairment, and severe mood disorders.1-4 Patients with advanced cancer who have dehydration or decreased oral intake almost always receive parenteral hydration in acute care facilities but almost never in hospices.3-5 In the past decade, the consequences of dehydration in terminally ill cancer patients have generated strong debate, with arguments for and against fluid administration.5-7
The decision of whether to administer fluids should be individualized, based on a careful assessment of a patient's clinical presentation, the potential advantages of parenteral fluids, and the patient's and family's wishes.6-8 Unfortunately, the decision-making process is complicated by the absence of randomized controlled trials focusing on the potential advantages and disadvantages of parenteral hydration.9 However, retrospective studies suggest that hydration can reduce neuropsychiatric symptoms such as sedation, hallucinations, myoclonus, and agitation.10,11
Methods of fluid hydration for palliative care patients include subcutaneous hydration or hypodermoclysis.6,12-14 This method has many potential advantages over other methods of fluid replacement: it may be started and stopped with no risk of thrombosis or bleeding; it is easier to manage in the home setting, with administration performed by family members or the patients themselves after minimal training; infusions can be easily administered by gravity, thereby avoiding the need for infusion pumps; and the same infusion site can be used for many days.6,13
The purpose of this randomized, controlled, double-blind study was to determine the effects of parenteral hydration with 1,000 mL/d versus 100 mL/d of normal saline, administered intravenously or subcutaneously, on overall symptom control in patients with advanced cancer.
PATIENTS AND METHODS
Study Design
This study was conducted at The University of Texas M.D. Anderson Cancer Center, Houston, TX; Hospital Eva Peron, Rosario, Argentina; Instituto Nacional de Cancer, Santiago, Chile; Clinica de Dolor, Fundacion Santafe de Bogotá, Bogotá, Columbia; and Centro Regional de Medicina, Los Montalvos, Spain. Patients had to meet the following inclusion criteria to be admitted to the study: a diagnosis of advanced cancer, defined as locally recurrent or metastatic, with no further treatment planned; an oral intake of less than 1,000 mL/d, as determined by clinical assessment; and evidence of mild to moderate dehydration, exhibited by decreased turgor in the subclavicular region lasting more than 2 seconds. In addition, patients had to have one or more of the following findings: dry mouth; thirst; decreased volume of urine output, as reported by the patient; a darker color of urine than usual, in the absence of reasons for jaundice or hematuria; and laboratory values consistent with dehydration, such as an elevated blood urea nitrogen to creatinine ratio of more than 20:1, when this value was obtained within 24 hours of admission to the study. Finally, patients had to be older than 16 years, able to understand and give consent for participation in the study, and able to tolerate parenteral treatment and the application of a subcutaneous or intravenous cannula. Exclusion criteria included a patient's refusal to participate; the presence of severe dehydration, defined as a decreased systolic resting blood pressure of 30 mmHg or lower from the patient's baseline value; low perfusion of the limbs; no urine output for 12 hours or longer; a decreased level of consciousness; or evidence of severe renal failure or bilateral hydronephrosis.
The study was carried out after institutional review boards at each study site approved the study and patients gave written informed consent to participate. All participating investigators underwent a 2-day training session at M.D. Anderson Cancer Center, during which the protocol was reviewed, including the method of administering hydration and the appropriate blinding of the type of parenteral infusion given. Random numbers determining treatment allocation were calculated at M.D. Anderson Cancer Center and delivered to the investigators in sealed envelopes. The randomization code was kept confidential at M.D. Anderson Cancer Center until the completion of the study.
The study was randomized, controlled, and double blind. Patients who met the inclusion criteria and agreed to participate were randomly assigned to receive either 1,000 mL of normal saline (treatment group) or 100 mL of normal saline (placebo group) as an infusion over 4 hours for 2 days. Both types of infusion were given parenterally (either intravenously or subcutaneously). Patients who already had an intravenous access device (n = 12) received infusions intravenously, and patients with no intravenous access device received infusions subcutaneously (n = 37). One investigator at each institution was unblinded and was responsible for preparing the saline, the droppers, and the portable and nonportable pumps to maintain the blinding. The investigator responsible for assessing the patient and the patient himself or herself were not aware of the amount of fluid being administered. In all cases, at least one test of blinding was conducted before activation of the study at each institution.
Outcome Measures
No previously defined outcomes for our study existed because no previous randomized controlled trials of hydration had occurred. We chose four target symptoms for hydration (sedation, fatigue, hallucinations, and myoclonus) on the basis of retrospective studies10-12,14,15 and our clinical experience. These target symptoms were found to be appropriate at a workshop at M.D. Anderson that included the study investigators and six other nonparticipating palliative care researchers from Houston. Patients scored the presence of the four target symptoms on a numeric scale of 0 (absent) to 10 (worst possible). The symptoms were considered present whenever the score was 1 or greater, and improvement was defined as a decrease of one point or more. The final scores were then added for the treatment (1,000 mL/d) and the placebo (100 mL/d) groups.
The main outcome of the study for the purpose of sample size calculation was the global assessment of the overall benefit of hydration to the patient, as determined by the physician and patient on day 2. The study was designed to detect a difference in the proportion of patients in the two groups having any important overall benefit. Parenteral hydration with 1,000 mL of normal saline was considered successful if 75% of the patients were perceived as having any important benefit (ie, a score of 3 or more) at the end of 2 days of hydration. We expected that approximately 50% of the patients receiving only 100 mL of normal saline would also be perceived as having an important benefit.
Our second proposed method of analysis was to test the two groups separately to determine whether the proportion of patients perceived to have some benefit was equal to 50% or greater than 50%.
The planned sample size was calculated as 54 patients per group to allow for the detection of differences, with an 80% power and a one-sided significance level of .05.
Assessments
The following six assessments were performed before parenteral hydration began (baseline) and at the end of days 1 and 2, unless otherwise noted: (1) a history, physical examination, and review of medications at baseline; (2) a Mini-Mental State Examination for assessment of cognitive function16 at baseline and on day 2; (3) an evaluation for target symptoms of dehydration, including hallucinations, myoclonus, fatigue, and sedation, during the previous 24 hours was assessed by an investigator who was unaware of the type of treatment that the patient received; these four symptoms were assessed using a numeric scale of 0 (no hallucination, no fatigue, and so on) to 10 (worst possible hallucinations, worst possible fatigue, and so on); (4) an assessment of the patient's well-being performed by both the patient and investigator at baseline and on day 2 using a numeric scale of 0 (worst possible) to 10 (best possible); (5) a global assessment of overall benefit on day 2 performed by the patient and investigator using a numeric scale of 1 (no important benefit) to 7 (greatly important benefit); (6) an assessment for toxicity (eg, pain, swelling, or leakage at the site of infusion) in each patient who received subcutaneous hydration using a numeric scale of 0 (no toxicity) to 10 (worst toxicity).
After the assessments on day 2, the physician and patient were free to choose to continue their preferred method of hydration. The brief 2-day period of observation was established to allow the treating physicians to provide their preferred method of hydration as rapidly as possible, after a sufficient time had passed for clinically detectable changes to occur. If, in the opinion of the treating physician, a patient experienced acute deterioration during the 2-day period, the patient was removed from the study and provided standard care.
RESULTS
The planned sample size of this study was 108 patients. However, because of difficulties in identifying patients with dehydration who had normal cognition and were willing to participate in the study, the total accrual was only 51 patients at the time of study closure. Twenty-eight patients were randomly assigned to the treatment group and 23 patients were randomly assigned to the placebo group (Fig 1). Only 49 (96%) of these patients were assessable; one patient from each group did not complete the study because of family refusal in one case and dyspnea and rapid deterioration in the other.
Table 1 lists the characteristics of the patients admitted to the study. All patients were receiving opioid analgesic agents for pain management.
Table 2 lists the effects of hydration on the target symptoms. Fifty-three (73%) of 73 target symptoms experienced by the treatment group improved, compared with 33 (49%) of 67 target symptoms in the placebo group (P = .006).
Patients blindly perceived hydration as effective in 17 (63%) of 27 cases in the treatment group and in nine (41%) of 22 cases (P = .78) in the placebo group. Investigators, who were unaware of the type of hydration patients received, perceived hydration as effective in 20 (74%) of 27 cases in the treatment group and 12 (54%) of 22 cases (P = .15) in the placebo group.
According to our study design, if 75% of patients had any perceived important benefit at the end of 2 days of hydration with 1,000 mL of normal saline, the treatment would be considered successful. We achieved this goal, with 21 (78%) of 27 patients in the treatment group having a perceived important benefit of the treatment. However, 13 (59%) of 22 patients in the placebo group also had a perceived important benefit of the treatment, which was higher than the expected 50%. The difference in the proportion of patients in the two groups who had a perceived important benefit was not statistically significant (P = .16). We had a 42% power to declare these proportions significantly different with our reduced numbers of 27 and 22 patients per group. Notably, if we had enrolled the original 54 patients per group and had found the same proportions (78% v 59%), we also would have had a reduced power of 61% to declare the difference in proportions statistically significant.
Our second proposed method of analysis was to test the two groups separately to determine whether the proportion of patients perceived to have some benefit was equal to 50% or greater than 50%. The results in the treatment group caused us to reject the null hypothesis (P = .0035), whereas those in the placebo group did not (P = .20). In other words, the proportion of patients in the treatment group judged by the treating physicians to have received a benefit was greater than 50%, with a significance level of .0035. In contrast, the proportion of patients in the placebo group judged by the treating physicians to have received a benefit was not significantly different from 50%.
The mean score (on a scale of 1 to 7) given by the investigators for their overall perception of the importance of the treatment benefit was 4.5 ± 2.3 in the treatment group and 3.4 ± 2.4 in the placebo group (P = .13; Table 3). The mean score given by patients was 3.8 ± 2.2 in the treatment group and 3.6 ± 2 in the placebo group (P > .2). The mean scores for the perception of an overall sensation of well-being are also provided in Table 3.
With subcutaneous administration of saline, the mean intensities of pain at the injection site were 2.10 ± 2.95 and 1.75 ± 2.55 for the treatment (n = 20) and placebo17 group (P = .1), respectively; the mean scores for injection-site swelling were 0.82 ± 1.13 and 1.41 ± 1.66 (P = .64). There were no significant differences in Mini Mental State Examination score at baseline and day 2 between treatment and placebo group.
DISCUSSION
In this randomized, controlled, double-blind study, we compared the effects of hydration with either 1,000 or 100 mL of normal saline on target symptoms and overall perception of treatment effectiveness by patients and investigators after 2 days of treatment. To our knowledge, this is the first randomized controlled trial comparing hydration with placebo in patients with advanced cancer. Our results suggested that hydration was able to improve the combined target symptom score and decrease myoclonus and sedation in the treatment group compared with the placebo group. A trend toward a perception of overall benefit for patients in the treatment group was also observed. Unfortunately, these results are far from conclusive and need to be confirmed by other studies.
Patient enrollment in this study was much more difficult than we had anticipated. At the time of their clinical dehydration and/or decreased oral intake, patients had frequently already developed delirium and were therefore unable to give consent for participation in the study. Future studies should attempt to obtain consent for participation before this complication occurs or use a design that allows proxy consent in patients who have already developed delirium.
Our findings suggested that double-blind studies are feasible and that patients identified for such studies are generally willing to give consent. One of the major challenges for such studies will be to obtain appropriate funding for conducting these studies, which are unlikely to be of interest to the pharmaceutical industry and so will have to compete for the limited funding allocated by granting agencies for clinical trials in palliative care.17
Some investigators were concerned about the possibility of clinical deterioration in the placebo group in a study conducted over multiple days. Patients, without receiving hydration, are likely to develop neurologic, renal, and cardiovascular complications as a result of the rapid decrease in intravascular volume. Clinicians who routinely administer parenteral hydration and believe there are clinical benefits from hydration have ethical concerns about randomly assigning patients to a placebo group. Future clinical trials of parenteral hydration should ideally be conducted in settings where hydration is not routinely administered, such as hospices. Therefore, the final assessment was conducted at the end of the infusion on day 2, which meant that the final assessment compared the differences between the treatment and placebo groups over a period of less than 36 hours. It is likely that much more significant differences will occur over time. Because the outcomes of this study were mostly of a neuropsychiatric nature, future studies should include more investigator assessments, given that delirium is an extremely frequent event in patients with advanced cancer and dehydration and that such patients may not be able to complete their own assessments.18 Family assessments regarding neuropsychiatric symptoms may be an important outcome as patients become progressively unresponsive. The lack of changes in cognition as measured by the Mini Mental State Examination is not surprising because normal cognition was a criterion for eligibility in these patients, and therefore there was very limited room for improvement. The outcomes chosen for this study were those found to be influenced by hydration in previous retrospective studies.10-12,14 Dry mouth and thirst were not included because previous research has shown that the association between these symptoms and the hydration status of patients with cancer is limited.19,20
The fact that 59% of patients in the placebo group perceived important overall symptomatic benefit after less than 36 hours emphasizes the need for double-blind studies. Our group has observed a large placebo effect in other symptom control studies in advanced cancer patients.21-24 It is possible that frequent contact with investigators results in improved symptom expression. Findings of this study and our previous experience21-24 suggest that a placebo control is very important even for preliminary symptom control studies. Our findings suggested that this blind perception of overall effectiveness might not be sensitive enough to detect the effects of full hydration compared with specific symptom assessments. The most effective outcomes will need to be carefully defined in future clinical trials. Prospective pilot studies of rehydration conducted on an open basis may help better delineate target symptoms for future randomized clinical trials.
Dehydration can cause various severe symptoms that can be difficult to interpret in terminally ill cancer patients because similar symptoms can be caused by the cancer itself or by the drugs the patient is taking25; such symptoms include dry mouth, cognitive failure, myoclonus, hallucinations, hyperalgesia, and grand mal seizures.26,10 Dehydration can also result in fatigue, nausea, postural hypotension, fever with no underlying infectious process, increased risk for bedsores, and constipation.26-28 On the other hand, parenteral hydration can increase the risk for edema and respiratory distress and requires that patients remain attached to an infusion device and undergo venipunctures. In addition, intravenous hydration is costly and difficult to maintain at home and, therefore, frequently requires patients to remain in the hospital until death.4-6
Our study found that hydration resulted in rapid decreases in the level of sedation, and myoclonus and a trend toward a decreased level of hallucinations. These benefits may have resulted from the hydration per se or simply from an increased elimination of active opioid metabolites from our patients, all of who were receiving opioids for their pain.15,18,25 This issue can be addressed in future studies by stratifying patients according the use of opioid therapy. From a pragmatic perspective, however, given that more than 80% of terminally ill cancer patients receive opioids, future clinical trials should focus mostly on patients such as those in our study.
Our findings suggest that clinical symptom improvement can be observed with a volume of hydration of approximately 1,000 mL/d. Previous research has shown that patients treated by palliative care teams receive significantly smaller volumes of hydration compared with those admitted to cancer centers.29 Subcutaneous hydration with a total daily volume of approximately 1,000 mL can be easily achieved at home, can be administered by patients and family members independently,6,9,14 and—as our preliminary findings suggest—can result in measurable symptomatic improvement.
Few areas of clinical care involve such disparity in clinical practice, such passionate rhetoric, and such lack of evidence as hydration near the end of life.5-9,15,20,26 Appropriately powered, randomized, double-blind studies with longer follow-up periods and more investigator-measured clinical outcomes than we used are required to help resolve some of these issues.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported by the Brown Foundation, Houston, TX; and the Tobacco Settlement Foundation.
Presented in part at the 40th Annual Meeting of the American Society of Clinical Oncology, New Orleans, LA, June 5-8, 2004.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Ripamonti C, Gemlo BT: Methods for artificial feeding and drainage of the gastrointestinal tract, in Bruera E, Higginson I (eds): Cachexia-Anorexia in Cancer Patients. New York, NY, Oxford University Press, 1996, pp 141-157
Alexander HR, Norton JA: Pathophysiology of cancer cachexia, in Doyle D, Hanks G, MacDonald N (eds): Oxford Textbook of Palliative Medicine. New York, NY, Oxford University Press, 1993, pp 316-329
Bruera E, Miller L, McCallion J, et al: Cognitive failure in patients with terminal cancer: A prospective study. J Pain Symptom Manage 7:192-195, 1992
Coyle N, Adelhardt J, Foley KM, et al: Character of terminal illness in the advanced cancer patient: Pain and other symptoms during the last four weeks of life. J Pain Symptom Manage 5:83-93, 1990
Billings JA: Comfort measures for the terminally ill: Is dehydration painful J Am Geriatr Soc 33:808-810, 1985
Fainsinger RL, MacEachern T, Miller MJ, et al: The use of hypodermoclysis for re-hydration in terminally ill cancer patients. J Pain Symptom Manage 9:298-302, 1994
Dunphy K, Finlay L, Rathbone G, et al: Rehydration in palliative and terminal care: If not, why not Palliat Med 9:221-228, 1995
Dunlop RJ, Ellershaw JE, Baines M, et al: On withholding nutrition and hydration in the terminally ill: Has palliative medicine gone too far A reply. J Med Ethics 21:141-143, 1995
Bruera E, MacDonald N: To hydrate or not to hydrate: How should it be J Clin Oncol 18:1156-1158, 2000
Bruera E, Franco JJ, Maltoni M, et al: Changing pattern of agitated impaired mental status in patients with advanced cancer: Association with cognitive monitoring, hydration and opioid rotation. J Pain Symptom Manage 10:287-291, 1995
de Stoutz ND, Bruera E, Suarez-Almazor M: Opioid rotation (OR) for toxicity reduction in terminal cancer patients. J Pain Symptom Manage 10:378-384, 1995
Bruera E, Brenneis C, Michaud M, et al: Use of the subcutaneous route for the administration of narcotics in patients with cancer pain. Cancer 62:407-411, 1988
MacMillan K, Bruera E, Kuehn N, et al: A prospective comparison study between a butterfly needle and a Teflon cannula for subcutaneous narcotic administration. J Pain Symptom Manage 9:82-84, 1994
Fainsinger RL, Bruera E: Hypodermoclysis for symptom control versus the Edmonton Injector. J Palliat Care 7:5-8, 1991
Lawlor PG: Delirium and dehydration: Some fluid for thought Support Care Cancer 10:445-454, 2002
Folstein MF, Folstein SE, McHugh P: "Mini-Mental State": A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12:189-198, 1975
Field MJ, Cassel CK: Approaching Death: Improving Care at the End of Life. Washington, DC, National Academy Press, 1997
Lawlor PG, Gagnon B, Mancini IL, et al: Occurrence, causes, and outcome of delirium in patients with advanced cancer: A prospective study. Arch Intern Med 160:786-794, 2000
Lawlor PG, Nekolaichuk C, Gagnon B, et al: Clinical utility, factor analysis, and further validation of the memorial delirium assessment scale in patients with advanced cancer: Assessing delirium in advanced cancer. Cancer 88:2859-2867, 2000
Morita T, Tei Y, Tsunoda J, et al: Determinants of the sensation of thirst in terminally ill cancer patients. Support Care Cancer 9:177-186, 2001
Bruera E, Roca E, Cedaro L, et al: Action of oral methylprednisolone in terminal cancer patients: A prospective randomized double-blind study. Cancer Treat Rep 69:751-754, 1985
Eduardo E, Moyano J, Sala R, et al: Dexamethasone in addition to metoclopramide for chronic nausea in patients with advanced cancer: A randomized controlled trial. J Pain Symptom Manage 28:381-388, 2004
Bruera E, Strasser F, Palmer JL, et al: Effect of fish oil on appetite and other symptoms in patients with advanced cancer and anorexia/ cachexia: A double-blind, placebo-controlled study. J Clin Oncol 21:129-134, 2003
Bruera E, Chadwick S, Brenneis C, et al: Methylphenidate associated with narcotics for the treatment of cancer pain. Cancer Treat Rep 71:67-70, 1987
Steiner N, Bruera E: Methods of hydration in palliative care patients. J Palliat Care 14:6-13, 1998
Burge FI: Dehydration symptoms of palliative care cancer patients. J Pain Symptom Manage 8:454-464, 1993
de Conno F, Ventafridda V, Saita L: Skin problems in advanced and terminal cancer patients. J Pain Symptom Manage 6:247-256, 1991
Sykes N: Constipation and diarrhea, in Doyle D, Hanks G, MacDonald N (eds): Oxford Textbook of Palliative Medicine. New York, NY, Oxford University Press, 1993, pp 299-310
Bruera E, Belzile M, Watanabe S, et al: Volume of hydration in terminal cancer patients. Support Care Cancer 4:147-150, 1996(Eduardo Bruera, Raul Sala)