How Useful Is Docusate in Patients at Risk for Constipation? A Systematic Review of the Evidence in the Chronically Ill

Transcription

1 130 Journal of Pain and Symptom Management Vol. 19 No. 2 February 2000 Review Article How Useful Is Docusate in Patients at Risk for Constipation? A Systematic Review of the Evidence in the Chronically Ill Virginia Hurdon, MD, CCFP, Raymond Viola, MD, MSc, CCFP, and Cori Schroder, MD, CCFP University of Ottawa Institute of Palliative Care, Ottawa, Ontario, Canada Abstract The effectiveness of docusate for constipation has not been studied in the terminally ill. Controversy also exists concerning its effectiveness in the chronically ill. Because chronically ill patients and terminally ill patients have several risk factors for constipation in common, we undertook a systematic review of prospective controlled trials of oral docusate in the chronically ill to clarify the utility of this drug in populations with advanced disease. The data sources were Medline 1966 April 1997, CINAHL 1982 April 1997, Current Contents August 1996 April 1997, Cochrane Library, a hand search of Index Medicus , three palliative care journals, references in relevant articles and texts, and direct contact with experts. Prospective controlled trials evaluating oral docusate in humans with chronic illness and identifiable risk factors for, or preexisting, constipation were selected. Only materials abstracted in English or French were considered. Information was collected by two independent reviewers and included patient demographic data, study design, dose of docusate, outcomes of stool consistency, stool frequency, need for other laxatives, and assessment of methodologic and reporting quality. Of nine identified studies, four were eligible. These incorporated three different designs and sample sizes that ranged from 15 to 74. Quality assessment scores were low (range with a perfect score being 1.0). Three studies were flawed in blinding of treatment allocation and the use of co-interventions. All studies showed a small trend toward increased stool frequency on docusate. Because of significant clinical heterogeneity in the identified studies, pooled data analysis was not feasible. At present, the use of docusate for constipation in palliative care is based on inadequate experimental evidence. Randomized controlled trials with chronically ill patients and patients with advanced disease are needed to determine its role in prevention and treatment of constipation. J Pain Symptom Manage 2000;19: U.S. Cancer Pain Relief Committee, Key Words Constipation, dioctyl, systematic review, chronic disease, terminal care Address reprint requests to: Virginia Hurdon, MD, CCFP, 43 Bruyere St., Ottawa, ON, Canada K1N 5C8. Accepted for publication: March 24, Introduction Prevention and effective treatment of constipation in terminally ill patients enhances symptom control, quality of life, and compliance with medications such as opioid analgesics. Current practice in terminally ill patients at risk for constipation often involves prophylaxis with an agent such as docusate in combination with a stimulating laxative such as sennosides or bisacodyl. Docusate (clinically used in salt form) acts as a surfactant, allowing water to U.S. Cancer Pain Relief Committee, /00/$ see front matter Published by Elsevier, New York, New York PII S (99)

2 Vol. 19 No. 2 February 2000 Docusate in Chronic Illness 131 penetrate the stool mass, 1 and increases intestinal intraluminal water and sodium secretion. 2 It is considered a softening agent, with little or no stimulant capacity. 1 Docusate has long been used to treat and prevent constipation in geriatric and institutionalized chronically ill patients, but in recent years, controversy has emerged regarding its effectiveness in these populations. 1,3 While some centers have removed it from their geriatric formularies (personal communication, D. Chornook, pharmacist, Ottawa Civic Hospital Geriatric Assessment Unit, 1997), others continue to prescribe it. No studies have been published examining the effectiveness of docusate for constipation specifically in palliative care patients. However, several studies have been performed in acute and chronically ill patients. Chronically ill patients and palliative care patients share risk factors for constipation. These include decreased physical activity, use of constipating medications, fiber-deficient diets, and illnesses that are progressive and often severely incapacitating. These similarities make the controversy regarding the effectiveness of docusate relevant to both palliative care and chronic care. The need to establish the efficacy of a drug, such as docusate, in medically ill populations is compelling. These patients are often already overburdened with the task of taking multiple medications. Available docusate preparations in North America may be difficult to take because of the large size of capsules or disagreeable taste. Fixed dose combinations of docusate and sennosides are also available, but these, like many such preparations, lack the flexibility to optimize the dose of the individual components; as well, the combination products are more expensive and often require a prescription. Prevention of unnecessary polypharmacy is increasingly important in health care resource management, as the number of chronically ill and terminally ill persons, and the proportion of public resources required to provide adequate care for them, continues to increase. In recent years, concerns about side effects of docusate, such as hepatotoxicity and undesired increase in systemic absorption of several medications, have been documented. 4 A systematic review of the evidence of effectiveness of docusate in chronic illness would assist in determining its rational place in therapy for both chronically ill and terminally ill patients. Such a review was performed to determine what effect docusate has on stool consistency, frequency, and need for other laxatives in chronically ill patients at risk for constipation. An important underlying hypothesis of this study was that docusate has similar clinical effects when used alone or, as more commonly prescribed, in combination with a stimulant laxative. Methods Study Inclusion Criteria A study was included if it fulfilled all of the following criteria (determined a priori): a) participants were adult humans described as chronically ill or inpatients in a chronic care facility; b) subjects with preexisting constipation or risk factors for constipation were included; c) the oral administration of docusate sodium or docusate calcium was evaluated in a prospective controlled fashion; d) grouped results were presented; e) at least one of the following variables was measured: stool consistency, stool frequency, or the use of other laxatives; and f) a full report or an abstract (if not reported in more complete form elsewhere) was available as original research in print or electronic form in a biomedical journal, a published letter to the editor, or an unpublished study identified by clinical content experts. Only studies published in English or French were eligible (because of lack of resources to translate from other languages). Studies published after 1940 were sought. Study Exclusion Criteria Retrospective study designs, case series, and case reports were excluded. Studies evaluating only a compound medication containing docusate were also excluded because the effect of docusate could not be separately examined. Patient satisfaction measures, while important, were excluded a priori from the review because of their subjectivity and anticipated inconsistent measurement. Search Strategy The search strategy was devised and executed by a single reviewer. To identify any previously indexed reviews on this topic, the Cochrane Database of Systematic Reviews and

3 132 Hurdon et al. Vol. 19 No. 2 February 2000 DARE (Database of Abstracts and Reviews of Effectiveness) were searched (Issue 3) using the terms constipation, dioctyl, and docusate. The Cochrane controlled clinical trials register was searched using the same terms. Medline was searched from 1966 April 1997 using the terms constipation, docusate, and dioctyl both as subject headings and as text words, doxidan as a textword, and docusate via its Medline drug registry number. Floating subheadings for drug therapy and prevention and control were used to maximize potentially relevant study retrieval. The Cumulative Index to Nursing and Allied Health Literature (CINAHL) database from 1982 to April 1997 was searched using the same search terms. The electronic database of Current Contents from August 1996 to April 1997 inclusive was searched using the textwords docusate, constipation, chronic, and elderly, and the combined term dioctyl and constipation. Because many of the early studies on docusate were published before 1966, a hand search of Index Medicus from 1940 to 1965 inclusive was performed using the terms cathartics, constipation, docusate, and dioctyl ; volumes of Index Medicus from 1957 to 1959 were not available locally, so an index of comparable scope, the Current List of Medical Literature, was substituted using the same search terms. Only articles containing the name of the drug or titled a new treatment/drug were retrieved. A hand search of three major palliative care journals Palliative Medicine (Edward Arnold Publishers, London, UK, Vol. 1 #1 Vol. 10 #4), the European Journal of Palliative Care (Hayward Medical Communications Ltd., Oxon, UK, Vol. 1 #1 Vol. 3 #3), and Progress in Palliative Care (Leeds Medical Information, Leeds, UK, Vol. 1 #1 Vol. 4 #5) was performed to identify potentially relevant studies. These journals were chosen because, with the exception of more recent volumes of Palliative Medicine, they are not indexed in Medline and are representative of western European publishing in palliative care. Selected current editions of textbooks in palliative care, gastroenterology, and geriatrics were searched for reference lists. Reference lists of all retrieved papers were scanned to identify further potentially relevant studies. Unpublished literature retrieval was restricted to information obtained from clinical content experts, who were contacted in writing to request their assistance in identifying any other potentially relevant material. A more extensive search of unindexed literature was not performed because of the lack of reproducibility of such searches. Study Selection and Quality Assessment All potentially eligible studies were independently assessed by two unblinded reviewers for inclusion. Disagreements regarding eligibility were resolved by consensus. Quality assessment was carried out on eligible and marginally eligible studies by two independent unblinded reviewers using an instrument developed specifically for drug studies of various designs. 5 The instrument contains 24 items and scores fall between 0.0 and 1.0, where 1.0 is the best possible score. Disagreements on scoring of individual items were resolved by consensus. Data Extraction and Synthesis Data extraction was performed by two independent unblinded reviewers and disagreements were resolved by consensus. A form was designed for this purpose and included the following main headings: study identification, demographic data, quality score, methodology, results, conclusions, and comments. It was anticipated that the studies retrieved would be heterogeneous with regard to study design, methodologic rigor, population baseline risk of constipation, dose of docusate, and method of outcome measurement. In this case, a narrative synthesis would be used. Results Included Studies Four studies 6 9 were identified that met the inclusion criteria (Table 1). All of these studies were identified through the electronic databases, reference list scanning, or the hand search of printed indexes. No relevant studies were identified from textbooks, journal hand searching, or field experts. No other systematic reviews on this topic were found, although one systematic review on treatment of chronic constipation in adults was identified. 10 All of the eligible studies were randomized controlled trials: two used crossover designs (Cass et al., Castle et al.) evaluating docusate against placebo; a third (Goodman et al.) was a

4 Vol. 19 No. 2 February 2000 Docusate in Chronic Illness 133 Results (as reported by original investigators) Author, Journal Quality (max 1.0), No. of patients completing/ recruited Design and duration Setting and population Cass and Frederik American Journal of Gastroenterology 1956 Table 1 Included Studies Drug form and dose Standardized protocol for other laxatives Outcomes of interest Placebo 4.0 Docusate /100 Randomized doubleblind crossover; placebo 10 days, docusate 20 days, no washout period Chronic disease hospital; all had chronic functional constipation (not defined) Castle et al. Geriatrics 1991 Docusate sodium 60 mg OD No Stool consistency 1 watery to 5 extremely hard (means) Stool frequency (mean stools/pt /week) Frequency of other laxative use (means) Placebo 3.5 Docusate 4.55 Placebo not allowed Docusate 0.042/pt/week Usual care 3.15/pt/week Placebo 93% soft or normal Docusate 97% soft or normal Placebo 4.12/wk 0.60 Docusate 4.25/wk 0.59 Goodman et al. Journal of Chronic Diseases /22 Randomized doubleblind crossover; placebo and docusate each 3 weeks with 2-week washout period before all treatment periods Nursing home unit of VA medical center; all patients on some form of bowel medication prior to study Docusate calcium 240 mg BID Yes Stool consistency as soft, normal or hard Stool frequency (mean stools/pt/wk SD) Frequency of other laxative use Placebo three doses Docusate four doses Docusate /34 Randomized unblinded comparative trial; duration range days; control group did not receive placebo Hospital chronic medical service; prevalence of preexisting constipation not measured Fain et al. Southern Medical Journal 1978 Docusate sodium 100 mg BID Yes for treatment group; not allowed in control group Stool consistency measured as total mean stools/pt during study SE A) watery, liquid B) soft formed, normal C) watery, hard formed D) hard dry stool E) enema Stool frequency (stools/pt) Control Docusate Grp A Grp B Grp C Grp A 1.05 Grp B 0.92 Grp C /47 Time series with 3 randomized singleblind treatment arms; placebo 2 weeks, followed by treatment for 3 weeks; no washout periods Retirement center; all had chronic, functional constipation and were dependent on laxatives (not further defined) SD standard deviation; SE standard error. Group A docusate sodium 100 mg OD Group B docusate sodium 100 mg BID Group C docusate calcium 240 mg OD Yes during treatment period; not allowed during placebo phase Stool frequency (mean/ pt/week SD) Frequency of other laxative use (mean # of enemas or suppositories/wk) Placebo Grp A Grp B Grp C Grp A 1.38 Grp B 1.03 Grp C 1.38

5 134 Hurdon et al. Vol. 19 No. 2 February 2000 parallel group trial evaluating docusate against an untreated control group; and the fourth (Fain et al.) was a time series study with multiple treatment arms. Years of publication ranged from 1956 to The settings included a free-standing retirement center (Fain et al.), a chronic disease hospital (Cass and Frederik), the nursing home unit of a veterans medical center (Castle et al.), and a chronic medical service in a hospital (Goodman et al.). Sample sizes (analyzable patients) ranged from 15 to 74 patients. In three of the four studies, subjects were assumed by the investigators to have preexisting constipation (based on patient self-report or chart stool records), although no study defined constipation. In the remaining study (Goodman et al.), the prestudy prevalence of constipation in the study subjects was not evaluated. All of the trials reported random treatment allocation, but only Fain et al. specified the method of randomization (flip of a coin). The degree to which investigators and subjects were blinded to treatment allocation varied. In Cass et al., blinding of patients and staff was reported, however the duration of the placebo arm was half that of the treatment arm, and other laxatives were permitted during the docusate phase but not during the placebo phase. Castle et al. reported double-blind conditions, whereas Goodman et al. did not report any blinding. Fain et al. reported blinding of patients; however the three groups received different doses of docusate (not reported to have been provided in a blinded form), and the patients lived in a communal setting in which it would be difficult to prevent interpatient contact. As well, other laxative use was permitted during docusate treatment but not while receiving placebo. Exclusion criteria varied considerably across studies. Cass et al. did not report any exclusion criteria and Castle et al. excluded patients with no bowel movement in the previous 72 hours or watery stools in the previous 48 hours, intestinal obstruction, ileus, undiagnosed abdominal pain, current gastrointestinal (GI) bleeding, GI surgery within the previous 3 months, creatinine clearance less than 10 ml/min, or exclusive tube feedings. Goodman et al. excluded only those with acute spinal cord injury, acute myocardial infarction, or diarrhea; Fain et al. excluded those unable to give consent, those having obstructive disease, neoplasms or other disorders of the gastrointestinal tract making it difficult to assess therapeutic response, and those with recent exposure to any of the study drugs. Similarly, identification of covariates such as diet, activity level, use of constipating medications, and presence of specific diseases was unevenly reported; however, this was compensated by study designs using patients as their own controls in all but one study (Goodman et al.). Duration of treatment with docusate was quite consistent, ranging from 20 to 26 days. In contrast, the doses ranged eightfold, from 60 mg/day of docusate sodium (Cass et al.) to 480 mg/day of docusate calcium (Castle et al.). Dose escalation was not performed in any study. Only Castle et al. used washout periods to reduce contamination due to prior therapy. Although all of the studies measured the three principal outcomes of interest, only Cass et al. and Goodman et al. presented data for all outcomes. Three of the four studies also measured adverse effects: two studies (Cass et al., Fain et al.) evaluated subjective reports of side effects; one study (Fain et al.) used physical examination and biochemical tests before and after treatment to evaluate adverse effects. Results are shown in Table 1. For Cass et al., results for the docusate phase were reported for two consecutive 10-day periods; these have been averaged in the table. Each of the individual studies showed a small trend toward increased frequency of bowel movements with docusate. A similar trend toward improved stool consistency was apparent in the two studies that reported data on this outcome (Cass et al., Goodman et al.). The effect of docusate on the use of other laxatives was not evaluable in three of the four studies (Cass et al., Goodman et al., Fain et al.) because placebo/control groups were not allowed to use any laxatives. In Castle et al., other laxative use did not differ between the two groups. Marginally Eligible and Excluded Studies Characteristics of the study 11 considered marginally eligible are summarized in Table 2. In Hyland and Foran, the study subjects were hospital inpatients, but it could not be determined from the report whether these patients had any chronic illnesses. The age range, gen-

6 Vol. 19 No. 2 February 2000 Docusate in Chronic Illness 135 Table 2 Marginally Eligible Studies Outcomes of interest Results (means) Other laxatives allowed? Drug form and dose Quality (max 1.0), No. of patients completing study Design and duration Setting and population Author, Journal Docusate vs. placebo mean difference 1.0 stools/ week 0.29 SE Not stated Not stated Not stated Stool frequency Docusate sodium 100 mg TID Hospital; no demographic data except that patients were geriatric and had persistent preexisting constipation Randomized double-blind crossover, 4 weeks each phase; no washouts /40 Hyland and Foran Practitioner 1968 Stool consistency Other laxative use der proportions, and method of recruitment were not reported. Of the 34 patients who completed the study, no data were presented on 19 who were not consistently constipated (definition not supplied) on placebo. Four studies (Table 3) were excluded: one because it examined docusate only as part of a compound medication, 12 two because they were case series, 13,15 and one because it evaluated docusate in chronic constipation not associated with chronic illness. 14 All of the excluded studies were published between 1956 and 1959, shortly after docusate had been introduced into the North American prescription drug market. Discussion It is apparent from Table 1 that the eligible studies were of different designs. Lack of methodologic rigor is reflected in uniformly low quality scores. Allowing additional laxative use only during docusate treatment in three of the four studies, as well as inadequate investigator and/or subject blinding in most of the studies potentially biased the results in favor of the active treatment. Only one of the studies (Castle et al.) may meet modern proposed criteria for the conduct and reporting of a randomized clinical trial, 16 but its sample size was very small (15 patients). Heterogeneity between the studies was evident in the chemical forms and doses of docusate used, the baseline risk of constipation, and the outcomes measured, such as stool consistency and use of other laxatives. Given these differences and concerns regarding the ability of individual studies to limit bias to a reasonable degree, it was decided that pooled data analysis would not provide a clinically sensible overall summary. Author Table 3 Excluded Studies Year of publication Reason for exclusion Brusch 1959 Compound medication Spiesman and Malow 1956 Not chronically ill; case series Friedman 1956 Not chronically ill; case series Harris 1957 Not chronically ill; case series

7 136 Hurdon et al. Vol. 19 No. 2 February 2000 Although this review attempted a comprehensive search of the published literature, some selection bias may have been introduced by omitting a formal search of European biomedical literature databases, such as EMBASE. Translation service limitations constrained selection of identified articles to those with English or French abstracts; however, no articles were identified containing abstracts in other languages. Selection of a chronically ill population with identifiable risk factors for constipation was performed to simulate as closely as possible the relevant risk factors in terminally ill patients. However, palliative care patients have other risks for constipation which resemble those in geriatric and acutely ill populations, and these could conceivably affect response to docusate. This review reveals that treatment of constipation with docusate in chronically ill patients and terminally ill patients is based on inadequate experimental evidence. A rigorously designed, randomized, double-blind, controlled trial in a well-defined population measuring standardized and clinically relevant outcomes would be of great assistance in clarifying the effectiveness of docusate in the chronically ill. It would also be useful, although methodologically more complex, to perform such a study in palliative care patients. This population may have superimposed conditions (e.g., side effects of drug therapy, hypercalcemia, acute spinal cord compression) and a fluctuating course of illness that may result in terminally ill patients responding differently than the chronically ill to docusate. Given the high prevalence of constipation in palliative care patients, and the weakness of the evidence on which use of docusate is based, a study comparing usual practice (i.e., docusate plus stimulant) with stimulant use alone would help clarify the role of docusate in the management of terminally ill patients. Acknowledgments The authors gratefully acknowledge the assistance of Jay Lynch, MSc, for his review of the manuscript, and the invaluable and expert advice of David Moher, MSc. References 1. Brunton LL. Agents affecting gastrointestinal water flux and motility; emesis and antiemetics; bile acids and pancreatic enzymes. In: Hardman J, Gilman A, Limbird L, eds. Goodman and Gilman s The pharmacological basis of therapeutics. New York: McGraw-Hill, 1996: Donowitz M, Binder HJ. Effect of dioctyl sodium sulfosuccinate on colonic fluid and electrolyte movement. Gastroenterology 1975;69: Yakabowich M. Prescribe with care: the role of laxatives in the treatment of constipation. J Gerontol Nurs 1990;16: Gattuso JM, Kamm MA. Adverse effects of drugs used in the management of constipation and diarrhea. Drug Safety 1994;10: Cho MK, Bero LA. Instruments for assessing the quality of drug studies published in the medical literature. JAMA 1994;272: Cass LJ, Frederik WS. Doxinate in the treatment of constipation. Am J Gastroent 1956;26: Castle SC, Cantrell M, Israel DS, Samuelson MJ. Constipation prevention: empiric use of stool softeners questioned. Geriatrics 1991;46: Fain AM, Susat R, Herring M, Dorton K. Treatment of constipation in geriatric and chronically ill patients: a comparison. Southern Med J 1978;71: Goodman J, Pang J, Bessman AN. Dioctyl sodium sulphosuccinate an ineffective prophylactic laxative. J Chron Dis 1976;29: Tramonte SM, Brand MB, Mulrow CD, Amato MG, O Keefe ME, Ramirez G. The treatment of chronic constipation in adults: a systematic review. J Gen Intern Med 1997;12: Hyland CM, Foran JD. Dioctyl sodium sulphosuccinate as a laxative in the elderly. Practitioner 1968;200: Brusch CA. Treatment of chronic functional constipation with dioctyl sodium sulphosuccinate and Danthron. J Ind Med Assoc 1959;10: Spiesman MG, Malow L. New fecal softener (Doxinate) in the treatment of constipation. Lancet 1956;76: Friedman M. Dioctyl sodium sulphosuccinate (Doxinate) in chronic functional constipation. Am Pract Dig Treat 1956;7: Harris R. Constipation in geriatrics: treatment with dioctyl sodium sulphosuccinate. Am J Dig Dis 1957;2: Begg C, et al. Improving the quality of reporting of randomized controlled trials: the CONSORT statement. JAMA 1996;276: