- © 2004 by American Society of Clinical Oncology
An Eicosapentaenoic Acid Supplement Versus Megestrol Acetate Versus Both for Patients With Cancer-Associated Wasting: A North Central Cancer Treatment Group and National Cancer Institute of Canada Collaborative Effort
- Aminah Jatoi,
- Kendrith Rowland,
- Charles L. Loprinzi,
- Jeff A. Sloan,
- Shaker R. Dakhil,
- Neil MacDonald,
- Bruno Gagnon,
- Paul J. Novotny,
- James A. Mailliard,
- Teresita I.L. Bushey,
- Suresh Nair and
- Brad Christensen
- From the Mayo Clinic and Foundation, Rochester; Duluth Community Clinical Oncology Program, Duluth, MN; Carle Cancer Center Community Clinical Oncology Program, Urbana, IL; Cancer Center of Kansas—Medical Arts Tower, Wichita, KS; Missouri Valley Cancer Consortium, Omaha, NE; Geisinger Clinic & Medical Center Community Clinical Oncology Program, Danville, PA; and McGill University, Montreal, Canada
- Address reprint requests to Aminah Jatoi, MD, Mayo Clinic, 200 First St SW, Rochester, MN 55905; e-mail: jatoi.aminah{at}mayo.edu
Abstract
Purpose Studies suggest eicosapentaenoic acid (EPA), an omega-3 fatty acid, augments weight, appetite, and survival in cancer-associated wasting. This study determined whether an EPA supplement—administered alone or with megestrol acetate (MA)—was more effective than MA.
Patients and Methods Four hundred twenty-one assessable patients with cancer-associated wasting were randomly assigned to an EPA supplement 1.09 g administered bid plus placebo; MA liquid suspension 600 mg/d plus an isocaloric, isonitrogenous supplement administered twice a day; or both. Eligible patients reported a 5-lb, 2-month weight loss and/or intake of less than 20 calories/kg/d.
Results A smaller percentage taking the EPA supplement gained ≥ 10% of baseline weight compared with those taking MA: 6% v 18%, respectively (P = .004). Combination therapy resulted in weight gain of ≥ 10% in 11% of patients (P = .17 across all arms). The percentage of patients with appetite improvement (North Central Cancer Treatment Group Questionnaire) was not statistically different: 63%, 69%, and 66%, in EPA-, MA-, and combination-treated arms, respectively (P = .69). In contrast, 4-week Functional Assessment of Anorexia/Cachexia Therapy scores suggested MA-containing arms experienced superior appetite stimulation compared with the EPA arm, with scores of 40, 55, and 55 in EPA-, MA-, and combination-treated arms, respectively (P = .004). Survival was not significantly different among arms. Global quality of life was not significantly different among groups. With the exception of increased impotence in MA-treated patients, toxicity was comparable.
Conclusion This EPA supplement, either alone or in combination with MA, does not improve weight or appetite better than MA alone.
INTRODUCTION
Bluefish, swordfish, salmon, and mackerel are rich in eicosapentaenoic acid (EPA), an essential omega-3 fatty acid, distinguished from other long-chain polyunsaturated fatty acids by a double bond that sits three carbons from the N-terminal of the molecule.1 Steadily amassing therapeutic promise, EPA has been studied in patients with heart disease, inflammatory bowel disease, psychiatric disorders,2-4 and in cancer patients who suffer from the syndrome of cancer-associated wasting.
Preliminary data on the latter have seemed promising. Loss of appetite and weight occur in the majority of oncology patients whose cancer has become refractory to antineoplastic therapy.5 To date, no intervention reverses the debility, deterioration in quality of life, and early demise characteristic of this pervasive syndrome in patients with refractory malignancies. Preliminary trials have suggested, however, that EPA might offer a promising approach. Wigmore et al6 observed weight stability with EPA (targeted dose 6 g/d) in weight-losing patients with pancreatic cancer. More importantly, Barber et al7 administered an EPA-containing dietary supplement that provided a total of 2.2 g of EPA/d to 20 weight-losing pancreatic cancer patients. At 3 weeks, they observed a median weight gain of 1 kg, along with an improvement in appetite and performance status. In yet another trial, Gogos et al8 randomly assigned 60 solid tumor patients to receive either a daily fish oil supplement that contained EPA (18 g of omega-3 polyunsaturated fatty acids plus vitamin E to compensate for the oxidative effect of the fish oil) versus placebo. Visual inspection of survival curves suggests a doubling of median survival in the EPA-supplemented arm compared with the placebo arm (P < .025). Taken together, these clinical data supported additional study of EPA for cancer-associated wasting.
In addition to these clinical trials, translational studies seem to provide the scientific underpinnings to explain a potential therapeutic role for EPA in cancer-associated wasting. First, EPA seems to suppress well-characterized mediators of cancer-associated wasting, including interleukin-6, an inflammatory cytokine.9 Four weeks of oral supplementation with EPA in pancreatic cancer patients resulted in a decrease in interleukin-6 in lipopolysaccharide-stimulated peripheral-blood mononuclear cells: 10.2 ± 2.1 v 3.5 ± 1.7 ng/mL (P < .05). Second, EPA seems to suppress the proteolysis-inducing factor, another well-described mediator.10 Using a mouse myoblast cell line, Smith et al11 observed that the proteolysis-inducing factor–induced protein degradation was eradicated with EPA pretreatment. Finally, EPA also seems to keep in check the ubiquitin-proteasome system, the pathway believed to be responsible for the bulk of muscle wasting in this syndrome.12 In animals bearing the MAC16 tumor, the proteasome became nonfunctional after EPA treatment.11 Thus, these translational studies drew the same conclusion as the clinical studies: EPA merited additional investigation.
This conclusion prompted this double-blinded, placebo-controlled trial, which compared the following three treatment arms: an EPA nutritional supplement plus placebo; megestrol acetate (MA) plus an isocaloric, isonitrogenous placebo; and both agents. The rationale for including MA in this study rests in the proven efficacy of this progestational agent in boosting appetite and nonfluid weight in cancer patients with this syndrome.13-15 The rationale for including a nutritional supplement, as opposed to EPA alone, comes from Barber's data7 that suggest caloric supplementation in combination with EPA leads to weight gain, not just weight stability. This multi-institutional, international clinical trial was therefore undertaken to determine whether an EPA-containing nutritional supplement—either alone or in combination with MA—improves weight, appetite, quality of life, and survival in advanced cancer patients who are suffering from cancer-associated wasting, compared with MA alone.
PATIENTS AND METHODS
Overview
The North Central Cancer Treatment Group (NCCTG) conducted this trial in collaboration with the National Cancer Institute of Canada. A total of 26 primary treatment sites participated. Institutional review boards at all institutions approved the study protocol, and patients provided informed written consent at enrollment.
Eligibility Criteria
Patients (≥ 18 years of age) with incurable malignancies, other than brain, breast, ovarian, prostate, or endometrial cancer, were eligible. Patients with these malignancies were excluded because either these tumors were potentially hormone-sensitive or because, as is the case with brain tumors, they are sometimes associated with persistent nausea and vomiting that might preclude consistent oral ingestion of the study agents. Patients had to have had an estimated life expectancy of ≥ 3 months and an Eastern Cooperative Oncology Group performance status of 2 or better. In addition, patients were required to have a self-reported, 2-month weight loss of at least 5 lb (2.3 kg) and/or a physician-estimated caloric intake of less than 20 calories/kg of body weight/d. Finally, all patients had to perceive loss of weight and/or appetite as a problem, and physicians had to view weight gain as beneficial. Chemotherapy or radiation was permitted.
Exclusion Criteria
Exclusion criteria included ongoing tube feedings or parenteral nutrition; edema or ascites; use of adrenal steroids (except short-term dexamethasone with chemotherapy), androgens, progestational agents, or other appetite stimulants within the past month; brain metastases; insulin-requiring diabetes; pregnant, nursing, or unwillingness to use contraceptives, if of child-bearing capacity; poorly controlled hypertension or congestive heart failure; history of thromboembolism; and obstruction of the alimentary tract, malabsorption, or intractable vomiting.
Stratification and Randomization
Patients were stratified based on the following: cancer type, gastrointestinal versus thoracic versus other; severity of weight loss, less than 10 lb (4.6 kg) v ≥ 10 lb in the preceding 2 months; planned concurrent chemotherapy, yes versus no; age, less than 50 v ≥ 50 years; the Good, Bad, Unsure (GBU) index (a validated16 stratification factor that incorporates Eastern Cooperative Oncology Group performance status, patient-completed Karnofsky score, physician-estimate of survival, and a patient-reported appetite); and medical center of enrollment. The stratification process used here was a minimization algorithm that balances the marginal distributions.17
Thereafter, in a double-blind manner, patients were randomly assigned to one of the following: an EPA supplement, two cans/d, plus a placebo liquid suspension that seemed identical to MA; MA liquid suspension 600 mg/d orally plus isocaloric, isonitrogenous placebo cans that seemed identical to the EPA supplement; or an EPA supplement two cans/d plus MA liquid suspension 600 mg/d orally, in combination.
In effect, this double-dummy study design used an active EPA supplement and an identical placebo. Both were isocaloric and isonitrogenous with 300 calories and 16 g protein in each can. Each EPA supplement can consisted of refined, deodorized sardine oil, which provided 1.09 g of EPA and 0.46 g of docosahexaenoic acid. The EPA supplement and the placebo also contained other nutrients including vitamins A, D, E, K, and C, along with folic acid and beta-carotene. These antioxidant nutrients were included to prevent the peroxidation of the highly polyunsaturated fatty acids.18 This EPA supplement was similar to that by Barber et al.7 In addition, given the prior experience with megestrol in the treatment of cancer-associated wasting,13 the testing of this agent in this trial was clearly justified.
Follow-Up
A history and physical examination, which included weight measurement in the oncologist's office, was performed at study entry and monthly. Toxicity assessment by the treating oncologist was completed during these visits.
Patients completed the NCCTG questionnaire for appetite and weight at baseline, weekly for 4 weeks, and then monthly. This questionnaire included a survey of toxicity, including nausea, vomiting, fluid retention, confusion, and impotence. A single-item Uniscale question measured global quality of life19 and was administered at the same frequency. The Functional Assessment of Anorexia/Cachexia (FAACT) questionnaire (version 4)20 was administered at the same frequency to assess appetite and appetite-related variables.
Patients continued treatment as long as both the patient and treating oncologist considered it beneficial or until concerning or intolerable side effects occurred.
Statistical Analyses
Each of the two alternative treatments were compared with the MA-alone arm by conducting two, two-sided tests using a 2% to 5% type I error rate. This approach results in an overall approximate error rate of a 5% chance of falsely concluding that at least one of the other two treatments is different from MA alone. The primary end point was an indicator variable of whether patients gained at least 10% nonfluid weight above baseline. Weight data were censored in patients who developed edema or ascites on physical examination or radiographic assessment. Patients rated their appetite with a previously validated NCCTG questionnaire.13 Differences between groups in these appetite-related categoric variables were analyzed with a Fisher's exact test.21 Continuous variables, such as quality-of-life ratings, ordinal baseline variables, and toxicity data, were compared among groups with Wilcoxon rank sum tests or independent sample t tests.22 Patients who dropped out before a weight assessment were considered to have experienced treatment failure in an intent-to-treat fashion.23 All analyses were carried out with two-sided tests and 2.5% type I error rates when using the reference arm approach as detailed above.
Sample Size Calculations
Primary power calculations are based on detecting differences in the proportion of patients who experience an a priori clinically significant outcome of 10% weight gain from baseline between each of the alternative treatments and the MA-alone arm. One hundred forty patients per arm enabled detection of a 10% difference in the percentage of patients who gained weight among the groups. This sample size allowed 80% power if the percentage improvement in the inferior group is no more than 5% of patients (two-sided Fisher's exact test with 2.5% type I error rate). It was also determined that 140 patients per arm allowed for 95% power to detect a shift in appetite improvement equivalent to one half of the standard deviation of the interval-level appetite scores. All sample size calculations took into account a 6-week median time on study, as is typical of previous clinical trials in advanced cancer patients with anorexia and/or weight loss.13-15
RESULTS
Baseline and Descriptive Data
A total of 429 patients were recruited between February 2000 and August 2002; 421 of these patients (98%) were deemed assessable. Patients were excluded from the analyses because of withdrawal before starting study drug (n = 6) or failure to meet eligibility criteria as determined after randomization (n = 2).
The three arms consisted of patient groups that were comparable at baseline on the basis of the stratification factors described earlier (Table 1). In addition, treatment arms were comparable with respect to baseline appetite, weight, and global quality-of-life scores. The latter was assessed by the single-item Uniscale (Table 1).
Baseline Patient Characteristics
The median number of days on study was relatively comparable between the EPA-treated, MA-treated, and combination-treated group (P = .71), and was slightly more than 3 months for the group as a whole. Patients withdrew for several reasons: declined further treatment and/or experienced toxicity (54%, 51%, and 48%); died (17%, 19%, and 16%); chose alternative treatment (1%, 4%, and 3%); developed other medical problems that precluded continued participation (12%, 8%, and 9%); and encountered “other” reasons, such as an inability to keep clinic appointments (16%, 17%, and 24%), with percentages in parentheses showing rates in the EPA-treated, MA-treated, and combination arms, respectively (P = .68 by χ2 test for the differences in distribution of reasons off study by treatment arm).
Weight
The primary end point was a 10% weight gain above baseline; a greater percentage of patients who received single-agent MA achieved this end point compared with those receiving only the EPA supplement. The percentage of patients who experienced such physician-reported weight gain is as follows: 6%, 18%, and 11% in EPA-treated, MA-treated, and combination-treated arms, respectively (P = .01; Table 2). Patient-reported weights, acquired from home scales, showed similar findings: 5%, 13%, and 7% in EPA-treated, MA-treated, and combination-treated arms, respectively (P = .08). Thus, the primary end point of this trial indicated greater efficacy with single-agent MA.
Maximum Weight Gain Over Baseline
When weight was explored with other analytic methods, findings were less consistent but continued to indicate increased efficacy with MA. Absolute weight gain, based on the last recorded physician weight, again showed that MA and combination therapy were superior. These analyses showed the following mean weight changes (in kilograms): −1.0, 1.3, and 0.1, in EPA-treated, MA-treated, and combination-treated arms, respectively (P = .008 for EPA v MA; P = .03 across all three arms). In contrast, when area under the curve (AUC) for weekly patient-reported weights was evaluated at 1 month, there were no statistically significant differences across all three arms (P = .38). Similarly, when the AUC for change in baseline weight for weekly patient-reported weights were evaluated at 1 month, there was no statistically significant difference over all three study arms (P = .13). However, a direct comparison of the AUC for change in baseline weight between the EPA-treated and MA-treated arms suggested a trend in favor of weight gain with MA (P = .05). Finally, when weight gain was evaluated with increments of less than 10% weight increase, there seemed to be no significant differences among the arms (Table 2). Although less conclusive overall, these secondary analyses nonetheless also suggest that single-agent MA is more effective in causing weight gain when viewed in aggregate.
Appetite
In contrast to the weight data, appetite results were less straightforward. Appetite, as measured by the NCCTG questionnaire, showed roughly comparable effects among groups. Summation of maximum favorable response questions (such as, “What effect, if any, do you feel the study medications have had on your food intake?”) showed favorable effects in all treatment arms, with the majority of patients describing notable but varying degrees of improvement above their baseline food intake: 64%, 68%, and 66%, in EPA-treated, MA-treated, and combination-therapy arms, respectively (P = .69). Table 3 lists detailed responses to appetite questions, including the one above; all results were relatively comparable among the three arms.
Percentage of Patients Reporting a Best Follow-Up Response to Select Appetite Questions
In contrast, the FAACT, which consists of questions with a strong emphasis on appetite, suggests appetite improved over time in the arms that included MA (Fig 1). For example, 4-week scores to rank the statement, “I have a good appetite,” suggested that MA and combination therapy provided better appetite stimulation compared with the EPA supplement alone, with scores of 40, 55, and 55 in EPA-treated, MA-treated, and combination-treated patients, respectively (P = .004).
Serial assessment of appetite with the Functional Assessment of Anorexia/Cachexia Therapy suggested that single-agent megestrol acetate provided better appetite stimulation compared with the eicosapentaenoic acid (EPA) supplement. Graph shows mean scores with 95% CIs.
Survival
There were no statistically significant differences in median survival in EPA-treated, MA-treated, and combination-treated arms, respectively (P = .82 by log-rank test; Fig 2).
There were no statistically significant differences in median survival cross all three study arms: (——) eicosapentaenoic acid supplement arm; (-- --) megestrol acetate arm; and ( · · · ) combination arm (P = .82 by log-rank test).
Quality of Life
With regard to global quality of life, the Uniscale detected no significant differences among maximally improved quality of life assessment over time in any of the three study arms. Maximum Uniscale scores over time minus baseline scores were as follows: median 0 (range, 0 to 71), 0 (range, 0 to 91), and 1 (range, 0 to 88) in EPA-treated, MA-treated, and combination-treated arms, respectively (P = .93, Wilcoxon rank sum).
Toxicity
Toxicity among the study arms was roughly comparable, and, for the most part, there were no severe or pervasive patient-reported or physician-reported toxicities that could be attributed specifically to one of the treatments tested in the study arms. The one exception was impotence. MA, a progestational agent, has been reported to cause male impotence, and this trial observed the same. Both the MA-treated and combination-treated male patients reported greater rates of impotence: 3% v 9% v 19%, as observed in EPA-treated, MA-treated, and combination-treated patients, respectively (P = .0006). It is important to point out, however, that such patient-reported symptoms were not assessed at baseline.
In addition to the impotence side effect, two other points relevant to toxicity deserve comment. First, in contrast to earlier studies with MA, in which thromboembolic phenomena were viewed as an adverse event, the incidence of such events in this trial occurred at roughly the same frequency: 6%, 8%, and 2% in EPA-treated, MA-treated, and combination-therapy arms, respectively. Second, prior studies suggest that MA mitigates nausea and vomiting, and this trial observed a comparable effect with the EPA nutritional supplement. The incidence of patient-reported nausea was as follows: 23%, 14%, and 16% in EPA-treated, MA-treated, and combination-treated arms, respectively (P = .11). Similarly, the incidence of patient-reported vomiting was as follows: 7%, 6%, and 9% in EPA-treated, MA-treated, and combination-treated arms, respectively (P = .52). Table 4 summarizes various patient-reported toxicities.
Patient-Reported Toxicities
DISCUSSION
This three-arm, multi-institutional trial included more than 400 patients with cancer-associated wasting and examined the following end points: weight, appetite, survival, and global quality of life. The primary end point of weight gain of 10% above baseline occurred in a greater percentage of patients who received MA alone compared with the EPA supplement (which contained 1.09 g of EPA administered bid), and combination therapy added nothing to single-agent MA. Appetite improvement, as assessed by the NCCTGquestionnaire, was roughly comparable in all three arms, although the FAACT seemed to suggest that the MA-treated arms gained greater appetite stimulation over time. Finally, survival and global quality of life were not statistically different among study arms. In short, this trial concludes that this EPA nutritional supplement was less effective than MA in causing a 10% weight gain above baseline, but that, depending on the method of measurement, it was relatively comparable to MA with respect to appetite stimulation, survival, and quality of life. Combination therapy did not add benefit above and beyond that provided by MA alone.
The data on appetite stimulation deserve further comment. As noted, findings varied depending in part on the instrument used. The NCCTG questionnaire suggested that appetite stimulation was roughly comparable among all three groups, whereas the FAACT suggested that MA provided greater appetite enhancement. In previous NCCTG comparative trials, these questionnaires have yielded far more consistent results and have clearly demonstrated the superiority of MA as an orexigenic agent compared with a variety of different agents, such as dronabinol, fluoxymesterone, and placebo.13-15 The fact that the NCCTG questionnaire did not find statistically significant differences in degree of appetite stimulation among treatment arms suggests that the EPA supplement might actually carry at least modest orexigenic effects. From a mechanistic standpoint, the ability of EPA to suppress cytokines that have been implicated in cancer-associated anorexia might in part explain such orexigenic effects.9
Although this trial found that MA is more effective than an EPA supplement in augmenting weight and perhaps might seem somewhat better in augmenting appetite, this progestational agent is nonetheless limited in its ability to treat cancer-associated wasting. Its clinical benefits are modest, with fewer than 30% of MA-treated patients experiencing short-term appetite stimulation independent of a placebo effect.13 Although weight and appetite improve, there is no improvement in survival or global quality of life.13-15 Hence, although MA may be successful in some respects, it is not highly successful in treating the syndrome of cancer-associated wasting. Indeed, the fact that this EPA supplement scored as well as MA on certain clinical end points (such as survival and global quality of life) only emphasizes the shortcomings of both these agents and the need to develop and test other approaches to treat cancer-associated wasting.
Previous studies had assessed EPA more favorably. These studies had suggested that EPA results in weight gain, appetite improvement, and a survival advantage. How might one explain the discordant findings between the previous studies and the study reported here? First, our trial did not include a pure placebo arm. Hence, MA might have masked some of the favorable effects of EPA. Had the present trial tested this EPA supplement against placebo, such orexigenic effects might have declared themselves more conspicuously by demonstrating a statistically significant difference among study arms. Second, although preliminary studies (such as those that preceded this trial) offer insight into promising therapies, the onus of supporting or refuting the clinical utility of such therapies falls to robust, multi-institutional, randomized trials. Our trial was developed and conducted with this responsibility in mind. Within this context, this EPA supplement seems to carry little therapeutic value for patients with cancer-associated wasting beyond that provided by MA. Along these same lines, Fearon et al24 recently provided preliminary data on a 200-patient trial in which an EPA nutritional supplement was compared with placebo in weight-losing pancreatic cancer patients. This trial, too, found no statistically significant difference in weight change between treatment groups at 4 and 8 weeks. Similarly, Bruera et al25 recently published a 60-patient trial with fish oil and concluded this agent was not effective. The scientific literature on cancer-associated wasting is replete with such large trials that have shifted the clinical and/or research focus from one agent only to allow opportunity for other agents or other approaches to be tested.14,26 This trial, along with those sited above, also seems to shift the focus. They suggest that other novel agents, perhaps tested as single agents or in combination with MA or EPA, merit additional investigation.
Finally, several points are pertinent to this study's design. First, this study did not include a placebo arm, as noted above. As a result, some potentially beneficial effects of this EPA supplement might have been overlooked. At the time of this study's development, however, at least one placebo-controlled trial was ongoing.24 Thus, the need for yet another controlled trial seemed unjustified in our opinion. Second, some might point out that a 10% weight gain is an ambitious end point for any agent aimed at treating cancer-associated wasting. However, it is important to emphasize that MA has resulted in this degree of weight gain in previous studies.13,15 Therefore, although it is ambitious, a 10% weight gain can certainly be justified as an achievable end point and was incorporated into this trial for this reason. Third, several other end points, such as fatigue and functional status, were not assessed in this study. The main goals of this study were to assess weight, appetite, quality of life, and survival. Our findings should not preclude future studies that might choose to evaluate these other end points. Finally, in terms of patient-reported side effects, it is important to point out that we did not assess baseline impotence and menses, and so definitive data to indicate change or lack of change over time are not available.
In summary, this study demonstrates that an EPA supplement does not result in notable improvement in weight, survival, or quality of life compared with MA. Combination therapy provides no benefit over and above MA alone. Additional study is needed to develop and test new agents and approaches for the treatment of cancer-associated wasting.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Footnotes
-
Conducted as a collaborative trial of the North Central Cancer Group, Mayo Clinic, and the National Cancer Institute of Canada, and supported in part by Public Health Service grants CA-25224, CA-37404, CA-15083, CA-63826, CA-63849, CA-35269, CA-35448, CA-35195, CA-35113, CA-60276, CA-52352, CA-35101, CA-35103, CA-63848, CA-35272, and CA-37417.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
- Received June 6, 2003.
- Accepted March 25, 2004.
