JOURNAL OF PALLIATIVE MEDICINE Volume 11, Number 4, 2008 © Mary Ann Liebert, Inc. DOI: 10.1089/jpm.2007.0178
A Comparison of Sennosides-Based Bowel Protocols with and without Docusate in Hospitalized Patients with Cancer PHILIPPA HELEN HAWLEY, B.Med., FRCPC1 and JAI JUN BYEON, M.D., Ph.D.2
ABSTRACT Background: Constipation is a common and distressing condition in patients with cancer, especially those taking opioid analgesics. Many institutions prevent and treat constipation with titrated laxatives, which is known as a bowel protocol. An effective and well-tolerated bowel protocol is a very important component of cancer care, and there is little evidence on which to base selection of the most appropriate agents. This study compares a protocol of the stimulant laxative sennosides alone with a protocol of sennosides plus the stool softener docusate, in hospitalized patients at an oncology center. The docusate-containing protocol had an initial docusate-only step for patients not taking opioids, and four to six 100-mg capsules of docusate sodium in addition to the sennosides for the rest of the protocol. Methods: Thirty patients received the sennosides-only (S) protocol and 30 the sennosides plus docusate (DS) protocol. The efficacy and adverse effects of the protocols were monitored for 5–12 days. The two protocols were used sequentially, creating two cohorts, one on each protocol. Eighty percent of patients were taking oral opioids and 72% were admitted for symptom control/supportive care. Results: Over a total of 488 days of observation it was found that the S protocol produced more bowel movements than the DS protocol, and in the symptom control/supportive care patients this difference was statistically significant (p 0.05). In the S group admitted for symptom control/supportive care 62.5% had a bowel movement more than 50% of days, as compared with 32% in those receiving the DS protocol. Fifty-seven percent of the DS group required additional interventions (lactulose, suppositories or enemas) compared to 40% in the S group. Cramps were reported equally by 3 (10%) patients in each group. Eight patients (27%) experienced diarrhea in the S group compared to 4 (13%) in the DS group. Conclusions: The addition of the initial docusate-only step and adding docusate 400–600 mg/d to the sennosides did not reduce bowel cramps, and was less effective in inducing laxation than the sennosides-only protocol. Further research into the appropriate use of docusate and into the details of bowel protocol design are required. INTRODUCTION
C
ONSTIPATION IS A FREQUENT SYMPTOM among hospitalized patients with cancer and is caused by many factors.1 Antiemetics, opioids, antidepressants,
and other constipating drugs are commonly used. Oral fluid intake is often severely reduced, and relative immobility is the norm, because people are tethered to infusion devices or are limited by symptoms such as fatigue, pain, and nausea. Constipation worsens nau-
1Department 2Division
of Palliative Care, BC Cancer Agency, Vancouver, British Columbia, Canada. of Palliative Care, University of British Columbia, Vancouver, British Columbia.
575
576
sea and anorexia, may cause discomfort, and increases the risk of delirium and urinary retention.2 Trauma from passage of hard stools and the need for rectal intervention may cause morbidity and even mortality, by providing a portal of entry for bacteria in an immunocompromised host. Many hospital units and home care programs have a “bowel protocol” for the prevention and treatment of constipation. When attempting to devise a bowel protocol there are very few studies from which to draw advice. There is no gold standard protocol. In hospice and palliative care units most patients (at least in Europe and North Americas programs) have cancer. The initiation of a bowel protocol for patients with cancer is often in an oncology center, and their protocol often continues to be used by patients as they progress to palliative and end-of-life care. Thus, an oncology center’s choice of bowel protocol may have an impact well beyond the acute oncology setting, and is an appropriate site in which to test the various agents available. In Canada, most oncology bowel protocols contain a stool softener such as docusate and a stimulant laxative such as sennosides. There is very limited evidence (but much clinical consensus) that sennosides are as effective as lactulose in the management of opioid-induced constipation.3 Sennosides are less expensive and more palatable than lactulose so are considered the laxative of choice in many institutions. Docusate is included in the belief that the softener may enhance the effectiveness of the stimulant laxative, and to reduce the incidence of cramps. Desiccation of stool is, however, thought to result primarily from a prolonged transit time in the colon rather than from abnormalities of mucosal water and electrolyte transportation mechanism.4 Docusate is the most commonly used bowel medication in the palliative care setting in North America5 but is rarely used in United Kingdom.6 There have been very few studies of docusate alone for constipation, and none in cancer patients on opioids.7–9 Only one study was randomized, and this was in only 22 long-term care home residents, and docusate 240 mg daily did not show any difference to placebo.9 The small number of other studies on docusate for constipation lack methodological rigor. In a systematic review of docusate in the chronically ill in 2000, Hurdon et al.8 concluded that the use of docusate for constipation in palliative care is based on inadequate experimental evidence. Even for the general population, the consensus of the American College of Gastroenterology Chronic Constipation Task Force in 2005 was that stool softeners have a minimal (if any) effect on chronic constipation.10 A 2007 review of opi-
HAWLEY AND BYEON
oid-induced constipation management by Ahmedzai and Burland11 included an updated Cochrane Review and also found no evidence to support the use of docusate. They also commented on the general lack of good evidence on which to base any laxative treatment recommendation. Docusate is relatively well-tolerated but is not completely free of adverse effects.12,13 It is advised to be taken 2 hours away from other medications to avoid disturbance to the absorption of the other medication. This can be difficult to schedule in patients taking multiple medications. Hospitalized patients with cancer are usually already heavily burdened by the need to take many pills. Docusate capsules are relatively large and a therapeutic dose requires patients to take 4 to 6 capsules per day. The hospital cost of purchasing docusate is very low (6¢–10¢ CDN per capsule) but the nursing workload for administration is significant. When the large number of doses administered is taken into consideration, docusate administration can represent a significant cost to an institution. Also, the cost to the patient is higher when the drug is purchased in a retail pharmacy (10¢–16¢ per capsule, or nearly $1 per day), and in many areas the cost of laxatives is not covered by health care plans. The purpose of this study was to compare the effectiveness of two sennosides-based bowel protocols, one with and one without docusate.
METHODS This was a nonrandomized, nonblinded sequential cohort study done in hospitalized patients with cancer. The study was approved by the British Columbia Cancer Agency (BCCA) ethics review board. Study subjects were recruited from the patients who were admitted to the Vancouver Center of the BCCA and who required a bowel protocol for at least 5 days to prevent or treat constipation. Patients were excluded if they wished to continue a previous bowel regimen, if they had a contraindication to docusate or sennosides, if they were unable to take oral medicine at any time during the study period, if they had colostomy or bowel obstruction, or if they were given a medication with laxative effect for any reason other than for constipation, for example lactulose for hepatic encephalopathy, oral magnesium for hypomagnesemia, or bowel preparation for endoscopy. The preexisting ward standard was a sennosidesplus-docusate bowel protocol, and initially eligible patients continued this protocol. Various methods of assessing effectiveness and tolerability were tested until
577
COMPARATIVE COHORT STUDY
it was felt that we could collect data in a meaningful way in all eligible patients. Because this study was unfunded and the investigators were not present on the ward on a continuous basis, it was necessary to devise a means for ward staff to recruit patients, obtain consent, and record the relevant data. Data recorded were those obtained routinely on all patients, and after some failed attempts to routinely gather patient satisfaction with their bowel management, no special assessments were able to be made for the study. Many different health care providers were involved in the care of each patient, and would have been unaware (after initial recruitment) that the patient was participating in the study. All staff were made aware that the study was taking place, and encouraged to document any bowel symptoms. Many patients were admitted to the ward during the study period that either did not require laxatives or who were not eligible for the study. Phase 1 consisted of the first 30 consecutive eligible patients using the docusate plus sennosides (DS) protocol. Phase 2 followed, during which the sennosides-only (S) protocol was introduced on the ward. Thirty consecutive eligible patients were recruited, hereinafter referred to as the sennosides-only (S) group. The starting dose was determined by the admitting physician based on whether or not they required opioids, and on their past bowel history. For those on opioids, or who had not had a bowel movement within the last 48 hours, the starting dose of sennosides would be the same with either protocol, but opioid-naïve patients without constipation would start with docusate alone on the combined protocol. The dose of docusate was initially 400 mg/d, increasing to 600 mg with step 2. Table 1 shows the two protocols. The sennosides preparation used was Senokot® (Purdue Pharma, Stamford, CT) and the generic docusate sodium capsules were made by Novopharm (Wilson, NC). Patients could take lactulose, a suppository, or an enema as needed, and use of these supplemental treatments was recorded. After patients were discharged, outcome data was collected by nursing chart review (J.J.B.). The narrative medical progress and nursing notes were also reviewed for any reporting of cramps or diarrhea. The primary outcome variables of this study were the proportion of total days with at least one bowel movement per day, and the proportion of patients with bowel movement at least 40% or at least 50% of days. The secondary outcome variables of the study were use of enemas, suppositories, or lactulose, and reported cramping and/or diarrhea. Diarrhea was defined as either passage of watery stool at least once per day, or more than 3 bowel movements a day. Age, gender, tu-
TABLE 1.
BOWEL PROTOCOLS
Docusate-plus-sennosides (DS) protocol a) Opiod naïve patients: start at step 1. b) Patients on opioid: if patient has not had a bowel movement in the 48 hours prior to assessment, start at step 3; otherwise start at step 2. c) For patients already taking laxative(s), switch to closest equivalent step of protocol. Step 1: Docusate sodium 200 mg po BID Step 2: Docusate sodium 200 mg po BID, sennosides 17.2 mg po q hs Step 3: If no bowel movement in 48 hours, docusate sodium 200 mg po TID, sennosides 17.2 mg po BID Step 4: If no bowel movement in the next 24 hours, docusate sodium 200 mg po TID, sennosides 17.2 mg po TID Step 5: If no bowel movement after a further 24 hours, docusate sodium 200 mg po TID, sennosides 25.8 mg po TID Sennosides-only (S) protocol a) Opiod naïve patients: start at step 1. b) Patients on opioid: if patient has not had a bowel movement in the 48 hours prior to assessment, start at step 2; otherwise start at step 1. c) For patients already taking laxative(s), switch to closest equivalent step of protocol. Step 1: Sennosides 17.2 mg po q hs Step 2: If no bowel movement in 48 hours, sennosides 17.2 mg po BID Step 3: If no bowel movement in the next 24 hours, sennosides 17.2 mg po TID Step 4: If no bowel movement after a further 24 hours, sennosides 25.8 mg po TID Docusate sodium 1 capsule 100 mg; sennosides 1 tablet 8.6 mg; PO, orally; BID, twice daily; TID, three times daily.
mor type, reason for admission, duration of days on bowel protocol, initial step on bowel protocol, and use of opioids were also collected. Use of opioids was defined as use of at least one dose of opioid per day throughout the time of use of the bowel protocol. If patients had been on the bowel protocol for more than 12 days, only the first 12 days were included for analysis, to balance the duration of observation in both groups. Subgroup analysis was done for patients who were admitted for symptom control or supportive care. We used Statistical Analysis System (SAS) 9.1 for statistical analysis (SAS Institute, Cary, NC). The t test was used to compare continuous variables, and 2 test and Fisher’s exact test were used to compare categorical variables. Multivariate analysis was done using multiple logistic regression. Age, gender, and reason for admission were adjusted in the model, because these were felt to be the factors with potential to affect bowel function.
578
HAWLEY AND BYEON TABLE 2.
GENERAL CHARACTERISTICS
Age (mean SD, range) Gender (number of male, %) Tumor type (number, %) Lung GI GU Breast Lymphoma/leukemia Head/neck/other Reason for admission (number, %) Supportive care or symptom control Diagnostic workup or chemotherapy Opioid use (number, %) Number of days on bowel protocol (mean, SD) Total days of observation (days) Initial step on bowel protocol (number, %) Step 1 Step 2 Step 3
OF
STUDY SUBJECTS
Docusate sennosides
Sennosides-only
p value
59.1 14.8 (24–87) 14 (46.7)
62.9 13.9 (25–85) 11 (36.7)
0.31 0.43
3 6 2 7 6 6
(10) (20) (6.7) (23.3) (20) (20)
4 3 14 3 3 3
(13.3) (10) (46.7) (10) (10) (10) 0.15
19 (63.3) 11 (36.7) 24 (80) 8.4 2.5 253
24 (80) 6 (20) 24 (80) 7.8 2.7 23
3 (10) 19 (63.3) 8 (26.7)
11 (36.7) 13 (43.3) 6 (20)
1 0.33 0.05
SD, standard deviation; GI, gastrointestinal; GU, genitourinary.
RESULTS Table 2 shows the demographics of the study subjects. Both groups were comparable in age, gender, and whether admitted for symptom control/supportive care or assessment/chemotherapy. Eighty percent of patients in both groups overall were taking opioids. In the 43 patients (72%) admitted for supportive care, 37 received opioids, 16 of whom were assigned to the DS protocol and 21 to the S protocol. Only those patients expected to stay longer than 5 days were recruited, and then only if the admitting resident or one of the other ward staff had time to obtain consent from the patient. Recruitment rates therefore differed between services and at different times, as residents rotated through their ward attachments. Genitourinary (GU) cancer was the most common diagnosis in the second cohort, the S group, whereas it was the least common diagnosis in the DS cohort.
TABLE 3.
RESULTS (ALL PATIENTS)
% of days with at least 1 BM/day (%) BM at least 50% of days, n (%) BM at least 40% of days, n (%) Use of lactulose, suppository, or enema, n (%) Diarrhea or cramping, n (%) BM, bowel movement.
This difference was due to the unexpected increase in enthusiasm of the GU team later in the study period. Patients in both groups were observed from 5 to 12 days, mean days of observation being 8 days in both groups. Ninety percent of patients in the DS group were receiving opioids, so were started on both docusate and sennosides, i.e., step 2, which had the same amount of sennosides as step 1 of the S protocol. Only 10% of patients in the DS group were started on docusate alone, (i.e., opioid-naïve and not already on laxatives or constipated). Twenty percent of the S group was started on level 3, whereas none of the DS group started at that protocol’s level 4, with an equivalent dose of sennosides. The dose of sennosides at the protocol starting level was therefore higher overall in the S group than in the DS group. Sixty-three percent of patients in the S group and 43% of the DS group had a bowel movement on at least 50% of days, but the difference was not statistically significant. In the subgroup analysis, for the pa-
Docusate sennosides
Sennosides only
p value
49% 13 (43.3) 18 (60).0 17 (56.7) 6 (20).0
50% 19 (63.3) 24 (80).0 12 (40).0 10 (33.3)
0.86 0.12 0.09 0.19 0.24
579
COMPARATIVE COHORT STUDY TABLE 4. FOR
RESULTS IN PATIENTS WHO WERE ADMITTED SYMPTOM CONTROL OR SUPPORTIVE CARE Docusate sennosides (n 19)
Sennosides-only (n 24)
p value
6 (31.6) 9 (47.4)
15 (62.5) 19 (79.2)
0.04 0.09
BM at least 50% of days, n (%) BM at least 40% of days, n (%) BM, bowel movement.
tients who were admitted for symptom control or supportive care, even with small numbers of patients a significantly higher number had bowel movement at least 50% of days in the S group. Looking at a lower level of effectiveness, 80% of the S group had a bowel movement at least 40% of days as compared with 60% of the DS group. This difference became statistically significant in the subgroup analysis for the patients who were admitted for symptom control or supportive care (Tables 3 and 4). When those who did not receive opioids were excluded from comparison, the S protocol gave better results (76% versus 50% of days) than the DS protocol, though the groups were even smaller and the difference did not reach statistical significance (Table 5). More patients in the DS group used extra interventions such as lactulose, suppositories, or enemas. More patients had diarrhea in the S group (8 compared to 4) but the number of patients who reported cramps was the same, 3 (10%) in both groups (Table 3). Diarrhea usually followed a period of no bowel movements. After adjusting for age, gender, and reason for admission by multiple logistic regression, the relative risk of having bowel movement more than 50% of days in the S group was 4.1 with a confidence interval of 1.2–14.8. The relative risk of having a bowel movement more than 40% of days in the S group was 6.4, with a confidence interval of 1.4–28.5.
DISCUSSION This small pilot study showed that the docusate plus sennosides bowel protocol was no more effective than
the sennosides-only protocol and did not decrease the risk of bowel cramping. There was a nonsignificant excess of patients experiencing diarrhea in the S group, however it is not clear from the clinical records if this was a desired outcome of achieving laxation, or whether it was considered an adverse effect by the patients. The sennosides-only protocol was shown to be significantly more effective in producing laxation than docusate-plus-sennosides protocol in those who were admitted for symptom control or supportive care, most of whom were also taking opioids. There are a number of possible reasons for this difference, including the possibility that docusate is an ineffective medication. Although the protocols were identical in the sennosides component, the DS protocol had an additional docusate-only step, which may have had the effect of delaying the sennosides dose escalation. There may be a degree of reluctance to start a protocol on step 2 because of fear of adverse effects from overdose of laxatives. The DS protocol also had a larger pill load, with 63% of patients in the DS group taking 6 pills per day at the initial step (4 docusate capsules and 2 sennosides tablets), and 27% starting with 10 pills per day (6 docusate and 4 sennosides). We speculate that patients and staff might be reluctant to start at an appropriate level of sennosides, or reluctant to move as instructed up the protocol, when the additional docusate pill load was included. The authors’ discussion with patients and staff suggests that this is an issue for many patients. A blinded study comparing combined sennosides/docusate pills versus sennosides-alone would be helpful in exploring the pill load effect further.
TABLE 5. RESULTS IN PATIENTS WHO WERE ADMITTED CONTROL OR SUPPORTIVE CARE AND WHO HAD BEEN
BM at least 50% of days, n (%) BM at least 40% of days, n (%) BM, bowel movement.
SYMPTOM OPIOIDS
FOR ON
Docusate sennosides (n 16)
Sennosides-only (n 21)
p value
5 (31.3) 8 (50.0)
12 (57.1) 16 (76.2)
0.12 0.10
580
This study has a number of significant weaknesses. Patients who wanted to continue a previous bowel protocol were excluded from the study. This might have excluded patients who had been taking docusate with satisfaction. However, the authors were not aware of any patients who declined to participate for this reason. When the authors visited the ward to promote recruitment, most patients not already on the study fulfilled the predetermined exclusion criteria. Those who were eligible but not already recruited were consented by one of the authors within 2 days of admission, the reason for delay usually being that the ward staff was either too busy or forgot, especially on weekends. This reflects the challenges of doing unfunded clinical research by a consultative service without full-time staff dedicated to the project. The two cohorts were not randomized and there were more genitourinary cancer patients recruited in the sennosides-only group. Five of the S patients received radiation to the spine or pelvis for bony metastasis/spinal cord compression (which might have caused laxation/diarrhea), compared to only two in the DS group. The constipating effect of antiemetics given during chemotherapy or radiotherapy might have also affected the results, however, more patients in the S group received chemotherapy or radiotherapy (6 radiotherapy, 8 chemotherapy, and 1 both) than in the DS group (2 radiotherapy and 8 chemotherapy, including 1 hepatic chemoembolization), which would have biased the study in favor of the DS protocol. Total patient numbers were small and the interventions were not blinded, however, many different staff were involved in documentation of bowel movements, which was recorded objectively. Another weakness was that as in any chart review, recording of side effects may not have been complete. Patients may have suffered adverse effects but not mentioned them to staff, or these effects may not have been recorded. Last, because this study took place in an acute oncology unit, the duration of observation was not sufficient to observe any differences in long term effectiveness, compliance, side effects or acceptability. A larger study will allow more comparable groups of patients and allow for statistically meaningful subgroup analysis. It is possible that different protocols will produce different results in different groups of patients. This pilot study has, however, given us preliminary evidence on which to base the selection of the sennosides-only protocol as our current standard. It provided us with useful information with which to design further studies in which stool consistency and overall satisfaction with a protocol will be assessed, as well as illustrating the need to study defined subgroups of pa-
HAWLEY AND BYEON
tients with respect to consumption of constipating medications, and administration of other treatments that might affect bowel function, such as specific chemotherapy or radiotherapy protocols. Our ward bowel assessment process and documentation methods are considered standard practice, but our experience with this study has shown that this may be inadequate for use in patients who are at high risk for constipation, and for whom the consequences of constipation can be very serious. Our research network is introducing a standardized assessment tool (Victoria Bowel Performance Scale, Victoria Hospice Society) into multiple palliative care settings throughout British Columbia, in conjunction with an educational initiative. This will ensure that good quality outcome assessment will be available as part of routine charting, and allow us to carry out better studies of constipation management in multiple care settings.
CONCLUSIONS A docusate-containing five-step bowel protocol was not as effective as a four-step sennosides-only protocol over a period of 5–12 days in a population of hospitalized cancer patients. The difference in effectiveness may have been because of lack of therapeutic benefit from docusate itself, differences in the format of the protocols, or because of differences between the patient groups. Further research is needed into the optimal laxative agents to use in this population at high risk of constipation and with the possibility of serious morbidity, and into the optimal protocol format. The lack of good evidence of benefit for docusate and the large quantities of this agent currently consumed in North America makes this a priority for further study. Meaningful interpretation of effects of laxatives in a clinical trial context requires more detailed bowel function assessment than is usual practice.
ACKNOWLEDGMENTS The authors would like to thank Juliana Man (pharmacist), for her assistance with preparation of the bowel protocols, and also all the patients, nurses, residents and clinical fellows who kindly made time out of their busy schedules to help with this study. Purdue Pharma has sponsored the provision of textbooks for palliative care residents, including J.J.B., in Canada through the Canadian Society of Palliative Care Physicians, and has sponsored a number of meetings
COMPARATIVE COHORT STUDY
through unrestricted educational grants to the BC Cancer Agency. Purdue Pharma did not participate in any way in this study and all study medications were purchased through the usual hospital purchasing process.
REFERENCES 1. Sykes N: Constipation and diarrhea. In: Doyle D, Hanks G, Cherny N, Calman K, (eds): Oxford Textbook of Palliative Medicine. New York: Oxford University Press, 2005, pp. 483–496. 2. McMillan SC: Presence and severity of constipation in hospice patients with advanced cancer. Am J Hosp Palliat Care 2002;9:426–430. 3. Agra Y, Sacristan A, Gonzalez M: Efficacy of senna versus lactulose in terminal cancer patients treated with opioids. J Pain Symptom Manage 1998;15:1–7. 4. Mancini I, Bruera E: Constipation in advanced cancer patients. Support Care Cancer 1998;6:356–364, 5. Curtis EB, Walsh TD: Prescribing practices of a palliative care service. J Pain Symptom Manage 1993;8:312–316. 6. Goodman M, Low J, Wilkinson S: Constipation management in palliative care: A survey of practices in the United Kingdom. J Pain Symptom Manage 2005;29:238–244. 7. Petticrew M, Watt I, Sheldon T: Systematic review of the effectiveness of laxatives in the elderly. Health Technol Assess 1997;1:i–iv, 1–52. 8. Hurdon V, Viola R, Schroder C: How useful is Docusate in patients at risk for constipation? A systematic review of the evidence in the chronically ill. J Pain Symptom Manage 2000;19:130–136.
581 9. Castle SC, Cantrell M, Israel DS, Samuelson MJ: Constipation prevention: Empiric use of stool softeners questioned. Geriatrics 1991;46:84–86. 10. Brandt LJ, Prather CM, Quigley EM, Schiller LR, Schoenfeld P, Talley NJ: Systematic review on the management of chronic constipation in North America. Am J Gastroenterol 2005;100(Suppl 1):S5–S21. 11. Ahmedzai SH, Boland J: Constipation in people prescribed opioids. BMJ Clin Evid 2007;6:2407 12. Gattuso JM, Kamm M: Adverse effects of drugs used in the management of constipation and diarrhea. Drug Safety 1994;10:47–65. 13. Thompson WG, Longstreth GF, Drossman DA, Heaton KW, Irvine EJ, Muller-Lissner SA: Functional bowel disorders and functional abdominal pain. Gut 1999;45(Suppl 2):II43–47. 14. Downing GM, Kuziemsky C, Lesperance M, Lau F, Syme A: Development and reliability testing of the Victoria Bowel Performance Scale (BPS). J Pain Symptom Manage 2007;34:513–522.
Address reprint requests to: Philippa Helen Hawley, B.Med. Department of Palliative Care BC Cancer Agency 600 West 10th Avenue Vancouver, British Columbia V5Z 4E6 Canada E-mail:
[email protected]
