Truth Seeker
New Member
Abstract
Background: A pilot study (NCT00316563) to determine if delta-9-tetrahydrocannabinol (THC) can improve taste and smell (chemosensory) perception as well as appetite, caloric intake, and quality of life (QOL) for cancer patients with chemosensory alterations.
Patients and methods: Adult advanced cancer patients, with poor appetite and chemosensory alterations, were recruited from two sites and randomized in a double-blinded manner to receive either THC (2.5 mg, Marinol®; Solvay Pharma Inc., n = 24) or placebo oral capsules (n = 22) twice daily for 18 days. Twenty-one patients completed the trial. At baseline and posttreatment, patients completed a panel of patient-reported outcomes: Taste and Smell Survey, 3-day food record, appetite and macronutrient preference assessments, QOL questionnaire, and an interview.
Results: THC and placebo groups were comparable at baseline. Compared with placebo, THC-treated patients reported improved (P = 0.026) and enhanced (P < 0.001) chemosensory perception and food 'tasted better' (P = 0.04). Premeal appetite (P = 0.05) and proportion of calories consumed as protein increased compared with placebo (P = 0.008). THC-treated patients reported increased quality of sleep (P = 0.025) and relaxation (P = 0.045). QOL scores and total caloric intake were improved in both THC and placebo groups.
Conclusions: THC may be useful in the palliation of chemosensory alterations and to improve food enjoyment for cancer patients.
introduction
Anorexia and weight loss contribute to functional loss, decreased survival, and poor quality of life (QOL) of advanced cancer patients [1]. The potential of delta-9-tetrahydrocannabinol (THC) to palliate these symptoms was suggested [2—8] and early work seemed to hold promise [2—5, 7]. Of seven studies [2—8], two investigated THC as an anti-emetic with the assessment of appetite as a side effect [2, 3], two were uncontrolled [4, 5], and one of two placebo-controlled studies used weight gain as an outcome following just 1 week of treatment [7]. In a randomized comparison with megestrol acetate [6], THC stimulated appetite in 50% of patients but was inferior to megestrol acetate. The results of a randomized trial by Strasser et al. [8] are difficult to interpret, owing to a large number of early deaths in all arms of the study.
THC increases appetite via endocannabinoid receptors (CB1r); appetite stimulation is documented in animals [9, 10] and in healthy human [11, 12] and acquired immunodeficiency syndrome (AIDS) populations [13, 14]. Other potential benefits of THC therapy have been overlooked. Chemosensory alterations are common and distressing among advanced cancer patients, and these contribute to decreased food intake and enjoyment and diminished QOL [15, 16]. Patients frequently report loss of food ideation and desire to eat [17, 18]. THC may increase food intake by stimulating the orosensory reward pathway, increasing motivation to eat energy dense foods and enhancing food enjoyment [10, 19]. CB1r are located in reward-related areas of the brain [20] and in the olfactory epithelium and bulb [20, 21] and are involved in peripheral odor processing [21] and potentially taste function [22]. We hypothesized that THC may favorably alter chemosensory perception. We therefore undertook a randomized placebo-controlled trial to determine THC's therapeutic potential to improve food-intake behavior for cancer patients with self-reported chemosensory alterations. Our approach included a panel of patient-reported outcomes related to food-intake behavior including perceived chemosensory alterations, macronutrient preference, caloric intake, appetite, and QOL. Safety and tolerability were also assessed.
patients and methods
This two-centre (Cross Cancer Institute, Edmonton, Alberta and Jewish General Hospital, Montreal, Quebec), phase II, randomized, double-blind, placebo-controlled 22-day pilot study (NCT00316563) was approved by Health Canada and the Research Ethics Boards of the Alberta Cancer Board, University of Alberta, and McGill University.
eligibility criteria
Adult patients with advanced cancer (defined as locally recurrent, locally advanced, or metastatic) of any site except brain who had a score ≥2 (out of 16) on a scored Taste and Smell Survey [23] were included (Table 1, see below). The survey identifies and quantifies chemosensory alterations; a score of ≥2 is associated with decreased caloric and protein intake and poorer QOL compared with no chemosensory alterations (T. Brisbois Clarkson, I. De Kock, S. Watanabe, V. Baracos, W. Wismer, unpublished data). Included patients had decreased food intake for ≥2 weeks (reported by subject or physician) and a physician-assessed life expectancy of >2 months. All patients spoke English and provided informed consent. Use of chemotherapy and radiation therapy was permitted during the trial, provided that no therapy-related adverse events (AEs) ensued. Eligibility of patients was determined by review of patient charts and verified by study physicians. Once approved, patients were enrolled by a research assistant.
exclusion criteria
Exclusion criteria included receiving enteral or parenteral nutrition; allergies or sensitivity to THC and/or sesame seed oil; history of substance abuse (determined by review of patients' medical records, alcohol abuse was often also assessed by CAGE questionnaire [24]) or psychotic episodes (e.g. diagnosis of schizophrenia or psychosis); mechanical obstruction of alimentary tract, mouth, or nose; radiation therapy to the head/neck; primary brain tumor; nausea score >5 on 11-point scale (0 = no nausea, 10 = worst possible nausea); medical conditions affecting chemosensory function (i.e. infection of the mouth or nasal cavity, active sinusitis, hay fever), history of tachyarrhythmias, angina pectoris, or uncontrolled hypertension; liver impairment determined by Child-Pugh score ≥10; use of marijuana within 30 days before start of trial.
Patients on treatments with the specific intention of increasing appetite or anabolism were excluded (e.g. dexamethasone 4-10 mg b.i.d., megestrol acetate, cannabinoids, oxandrolone). Low-dose corticosteroids for indications other than appetite were allowed provided that doses remained constant for the duration of the trial. Patients were screened for oral candidiasis; if present, patients were eligible to participate once the infection was successfully treated.
random treatment assignment, blinding, and intervention
After baseline assessments, patients were randomly assigned by a third party pharmacist in a double-blinded manner to receive either THC (Marinol®; dronabinol 2.5 mg capsules, Solvay Pharma Inc.) or placebo according to a third-party computer-generated randomization scheme. Patients started on THC 2.5 mg or placebo once daily for the first 3 days (before bedtime for first 2 days and before supper on third day). The dose was increased to THC 2.5 mg or placebo twice daily (1 capsule before lunch and supper) on the fourth day (Figure 1). Patients had the option to increase their drug dose to a maximum of 20 mg/day [25].
outcome measures
Patients completed assessments at baseline (day 0) and after 18 days of treatment with 1 day latitude in posttreatment assessments (Figure 1). All assessments used patient-reported outcomes to capture and describe changes experienced and valued by the patients [26—28].
The Taste and Smell Survey [23] was used to identify and quantify chemosensory alterations 'since study treatment'. This tool has a maximum chemosensory complaint score of 16 (Table 1) [16]. Alterations in taste and smell emerging during study treatment could be pleasant or unpleasant. For example, a loss of sensation resulting in food being perceived as tasteless would be unpleasant, whereas the loss of a heightened and distressing sensitivity to tastes or odors could be perceived as a benefit. We therefore added the qualifying statements 'yes, it's better', 'yes, it's worse', and 'no, it's the same' to questions 1—4; 'chemosensory improvement' was calculated by tallying the 'yes, it's better' responses to these questions. Responses to questions 8—12 were further qualified as to whether change was 'pleasant or unpleasant'; responses of 'pleasantly stronger' comprised the 'chemosensory enhancement' outcome (maximum score of 5). The survey also includes open-ended questions for subjects to elaborate on the specific nature of chemosensory change(s) and their impact on QOL.
The 100 mm Satiety Labeled Intensity Magnitude (SLIM) scale [29] was completed 10—15 min before each meal for 1 day pretreatment and following 18 days of treatment for an assessment of appetite. This scale is anchored with greatest imaginable fullness = 0 and greatest imaginable hunger = 100 (neither hungry nor full = 50). The Macronutrient Preference Checklist (MPC) [30] was completed with the SLIM to assess concurrent macronutrient preferences. The MPC is scored on the number of food items selected (0—8) in each of four macronutrient categories (high protein, high fat, high carbohydrate, low energy). SLIM and MPC premeal scores were averaged for an overall day score. A 3-day dietary record [31] was used to estimate total calories and macronutrient intake (Food Processor II Nutrient Analysis Program; Esha Research, Salem, OR).
QOL was assessed with the Functional Assessment of Anorexia/Cachexia Therapy (FAACT) questionnaire [32]. The 11-point Edmonton Symptom Assessment System [33] was used to assess nausea. A Side Effect Survey [34] documented the tolerability of the study drug. Interviews were conducted to determine patients' treatment-related changes in food preferences and chemosensory alterations.
patient characteristics
Patients were recruited from either the palliative home care program or outpatient clinics at local cancer clinics in Edmonton, Alberta, and Montreal, Quebec, Canada, over 2.5 years (2006—2008). There were no differences for any outcomes between study sites (P < 0.05). Patient characteristics (Table 2) and dropout rates (Figure 2) were similar for THC and placebo groups. Of factors that could affect chemosensory perception, 33% of patients were receiving chemotherapy at the time of data collection (although all patients had previously received multiple rounds of various therapies). Similar proportions of participants in each treatment arm reported wearing dentures, experiencing previous mouth or gum infections, and previous taste or smell problems and were previous or current smokers. No patients were taking concurrent appetite stimulants. All participants were treated with various prior chemotherapies (Table 2); however, their effect on chemosensation is impossible to deconvolute in a sample of this size with diverse cancer diagnoses.
statistical analyses
Statistical analyses were performed on a per protocol analysis basis [35] using SAS (Statistical Analysis System for Windows, Cary, NC) [36]. Total and subdomain scores of the Taste and Smell Survey were treated as the primary outcomes, with secondary outcomes of appetite, macronutrient intake, nausea, and QOL. In this exploratory study, the nature of treatment effects and potential effect size(s) were unknown and for that reason, 10 participants per arm were assessed. Descriptive statistics were used to describe prevalence, nature, and severity of chemosensory alterations. Chi-square and Fisher's exact tests were used to evaluate patient characteristics, yes/no responses, treatment side effects, and AEs. Time series analysis of variance [37] with baseline assessments as covariates were significant [38] and were used to assess differences in chemosensory complaint scores, caloric intake, appetite, macronutrient preferences, QOL, and nausea between- and within-treatment groups as these data met assumptions of normality. Pairwise differences of least squares means (pdiff) were used for post hoc comparisons.
results
In the THC group, eight patients followed the dosing protocol (i.e. 2.5 mg b.i.d.) and three patients increased to 2.5 mg t.i.d. by taking an additional 2.5 mg before supper. In the placebo group, seven patients followed the dosing protocol and three patients increased their dose to 3 capsules/day.
taste and smell perception — results of the scored questions
Total chemosensory complaint scores decreased with THC treatment compared with baseline but were not different from placebo (Table 3). Chemosensory enhancement (frequency of response 'pleasantly stronger' on questions 8—12; Table 1) was increased from baseline with THC treatment compared with placebo (P = 0.018) (Table 3). Chemosensory improvement (frequency of response 'yes, it's better' on questions 1—4; Table 1) was twice as frequent with THC treatment (36%) compared with placebo (15%) (P = 0.026).
taste and smell perception — responses to open-ended questions and interview
Responses to open-ended questions and interview were concordant with the outcomes described above. The majority (73%) of THC-treated patients reported an increased overall appreciation of food compared with patients receiving placebo (30%) and more often stated that study medication 'made food taste better' (55%) compared with placebo (10%) (P = 0.04). Half of the patients who reported odors to be unpleasant at baseline no longer found odors offensive with THC treatment (P = 0.083). The majority of THC-treated patients (73%) indicated a renewed ability to discriminate tastes, flavors, and food odors. By contrast, 80% of patients in the placebo group reported their taste and smell function to be the 'same as before' (60%) or 'worse' (20%) compared with baseline.
appetite
For the THC group, SLIM appetite scores increased relative to baseline and placebo (Table 3). The majority of THC-treated patients (64%) had increased appetite, three patients (27%) showed no change and one patient's data were incomplete. No THC-treated patients showed a decrease in appetite. By contrast, the majority of patients receiving placebo had either decreased appetite (50%) or showed no change (20%).
food preferences and caloric intake
Compared with placebo, THC-treated patients increased their protein intake as a proportion of total energy. There was a trend for THC-treated patients to express preference for high protein foods on the MPC (Table 3). During interview, patients in the THC group (55%) commonly reported savory foods (e.g. hamburgers, chicken, fish, baked beans, and mushrooms) to 'taste better' and to be 'more appealing' since the study treatment. No patients in the placebo group reported an increased liking of meats.
Total caloric intake (Table 3) or intake per kilogram body weight (P = 0.557) did not differ between-treatment groups. Relative to baseline, 73% of THC-treated patients increased their caloric intake (range 100—775 kcal/day) compared with 50% patients in the placebo group (100—965 kcal/day).
quality of life
FAACT global QOL scores improved similarly for both THC and placebo groups (Table 3). The FAACT subdomain of anorexia—cachexia-related nutritional well-being improved in the THC group but was not different from placebo.
side effects and AEs
Nausea scores were unaffected by THC treatment (P = 0.532). Quality of sleep and relaxation were more frequently reported to be 'pleasant' by THC-treated patients compared with placebo on the Side Effect Survey (Table 3). There were no other differences in survey responses between-treatment groups (P > 0.05; Table 4).
THC was well tolerated. No differences were reported during the trial or within the 30-day follow-up period between THC and placebo groups for the number of AEs or serious AE (SAE) (P = 0.622 and P = 0.244, respectively). Most AEs were unrelated to THC therapy, six were unclear, and four were possibly related (Table 5). Most SAE were also unrelated to THC therapy, four were unclear and one was possibly related to treatment (irregular heart beat).
discussion
main findings
Our pilot study demonstrates that THC, compared with placebo, improved and enhanced chemosensory perception, altered macronutrient preference, appeal of savory foods, appetite, relaxation, and quality of sleep for advanced cancer patients with chemosensory alterations.
We opted for self-reported chemosensory perception as the most relevant predictor of food preference and enjoyment in lieu of objective clinical measures (i.e. millimolar concentration thresholds for detection of individual tastants and odorants). Along with Bartoshuk, we speculate that taste and smell alterations are not merely quantifiable physiological changes but also involve the loss of food enjoyment [18]. Bernhardson et al. [39] and Steinbach et al. [40] have recently used self-reported chemosensory perception to comprehensively describe taste and smell changes experienced by cancer patients. Clinical measures of chemosensation cannot capture dimensions such as flavor, food enjoyment, or impact on patient's food-intake behavior. The millimolar concentration threshold for perception of salt dissolved in water is not a surrogate for the complex perceptions that contribute to the ingestion and enjoyment of food.
Our findings parallel earlier results of enhanced sensory perception and improved food enjoyment among healthy marijuana users [41, 42]. THC also reduces the unpleasantness of a bitter taste solution in animals [19]. Our THC-treated patients reported odors and the taste of meat to be less offensive, possibly contributing to increased calories ingested as protein. These results suggest that THC improved chemosensory perception through reward systems [42].
Our findings are important as there is no accepted treatment for chemosensory alterations experienced by cancer patients [15, 16]. THC treatment may hold multiple clinical benefits for cancer patients, beyond its indication as a treatment for nausea and its effects on appetite. THC may palliate an array of symptoms: chemosensory alterations, food enjoyment, pain, depression, anxiety, poor quality of sleep, and inflammation [43]. Food enjoyment is a component of QOL, and while food enjoyment is intangible in terms of tools for its quantification, we nonetheless suggest that THC may well contribute to the overall enjoyment of food in cancer patients. Improved sleep may be due to the presence of CB1r in the basal forebrain [20] or related to increased relaxation noted in various populations [7, 44]. Here, improved quality of sleep and relaxation may have increased appetite and improved chemosensory perception, encouraging a positive outlook on food [45].
prior and future clinical trials of THC
THC stimulates appetite in healthy volunteers [11, 12] and AIDS patients [13, 14], but its ability to do so in cancer patients has not been consistently reported. THC increased appetite for 34%—72% of cancer patients with doses ranging from 5 to 45 mg/day [3—8, 44]. Nelson et al. [4] showed promising THC appetite stimulation in cancer and 6 of 18 patients opted to remain on THC treatment for improved appetite and food intake [46]. However, this study was criticized for the lack of a control group. Jatoi et al. [6] reported THC to stimulate appetite in 50% of patients. Strasser et al. [8] noted no differences between THC or THC + cannabidiol and placebo for appetite or QOL. In our study, 64% of THC-treated patients showed improved SLIM appetite scores and this assessment was not susceptible to placebo effect. The verbally anchored SLIM scale shows greater sensitivity and specificity than unlabelled visual analog scales [29], which had been used in other studies [8].
Jatoi et al. [6] and Strasser et al. [8] selected 5 mg THC daily to minimize side effects; however, Nelson et al. [4] showed more promising results with 7.5 mg THC daily. AIDS patients tolerated doses as high as 40 mg THC daily well [13]. We started patients at a low dose to build-up tolerance and minimize negative psychoactive effects [25] and allowed patients to titrate their dose upward and this was well tolerated. We noted dropouts and withdrawn consents due to changes in health status unrelated to study treatment. The dropout rate was not uncommon for an advanced cancer population [6]. Clinical trials in advanced cancer have the added complexity of comorbidities and expected death. The exclusion of data that are confounded by poor prognosis is critical for interpretable results, which may be a criticism of previous work [8].
One consideration for future trials will be the situation of therapy for taste and smell alterations within the disease trajectory, as these problems have been characterized both during first-line therapy in treatment naïve patients [47] as well as in advanced stage patients with progressive disease who have received multiple lines and cycles of treatment, as here. These may be related or distinct clinical entities.
Our sample size and study duration within a placebo-controlled design showed statistical significance for several outcomes responsive to THC: chemosensory perception, appetite, relaxation, and quality of sleep. However, our results require verification in larger trials. Our goal was to conduct a proof of principle study to provide direction for future trials. The data here would assist in the development of larger phase II trials by facilitating sample size calculations given our preliminary indications of effect sizes and variance (e.g. total chemosensory complaint scores as the primary outcome would require ∼50—60 patients in each treatment group based on the variance determined here). Our results offer a reasonable starting point from which future studies may investigate the use of THC in cancer anorexia where chemosensory alterations are present.
For the design of future trials, it seems important to (i) include a placebo group as outcomes may appear more favorable when compared with drug alone; (ii) include patient-reported assessments to capture relevant aspects of food-intake behavior, such as chemosensory and food preference changes; and (iii) power studies around differentiable outcomes, such as chemosensory complaint scores. As absorption of oral THC varies greatly between individuals [25, 48], and given the controversy surrounding the appropriate dose in cancer, future trials may allow patients to titrate their dose or compare different dose levels. Inevitably, questions are raised about the ability to blind THC treatment based on its well-known psychoactive characteristics; however, the timed administration of low doses and the lack of differences in AEs between-treatment groups suggest that this problem was likely minimal. THC merits further investigation as a therapy for patients who suffer from chemosensory alterations and loss of food enjoyment.
Source, Graphs and Figures: Delta-9-tetrahydrocannabinol may palliate altered chemosensory perception in cancer patients: results of a randomized, double-blind, placebo-controlled pilot trial
Background: A pilot study (NCT00316563) to determine if delta-9-tetrahydrocannabinol (THC) can improve taste and smell (chemosensory) perception as well as appetite, caloric intake, and quality of life (QOL) for cancer patients with chemosensory alterations.
Patients and methods: Adult advanced cancer patients, with poor appetite and chemosensory alterations, were recruited from two sites and randomized in a double-blinded manner to receive either THC (2.5 mg, Marinol®; Solvay Pharma Inc., n = 24) or placebo oral capsules (n = 22) twice daily for 18 days. Twenty-one patients completed the trial. At baseline and posttreatment, patients completed a panel of patient-reported outcomes: Taste and Smell Survey, 3-day food record, appetite and macronutrient preference assessments, QOL questionnaire, and an interview.
Results: THC and placebo groups were comparable at baseline. Compared with placebo, THC-treated patients reported improved (P = 0.026) and enhanced (P < 0.001) chemosensory perception and food 'tasted better' (P = 0.04). Premeal appetite (P = 0.05) and proportion of calories consumed as protein increased compared with placebo (P = 0.008). THC-treated patients reported increased quality of sleep (P = 0.025) and relaxation (P = 0.045). QOL scores and total caloric intake were improved in both THC and placebo groups.
Conclusions: THC may be useful in the palliation of chemosensory alterations and to improve food enjoyment for cancer patients.
introduction
Anorexia and weight loss contribute to functional loss, decreased survival, and poor quality of life (QOL) of advanced cancer patients [1]. The potential of delta-9-tetrahydrocannabinol (THC) to palliate these symptoms was suggested [2—8] and early work seemed to hold promise [2—5, 7]. Of seven studies [2—8], two investigated THC as an anti-emetic with the assessment of appetite as a side effect [2, 3], two were uncontrolled [4, 5], and one of two placebo-controlled studies used weight gain as an outcome following just 1 week of treatment [7]. In a randomized comparison with megestrol acetate [6], THC stimulated appetite in 50% of patients but was inferior to megestrol acetate. The results of a randomized trial by Strasser et al. [8] are difficult to interpret, owing to a large number of early deaths in all arms of the study.
THC increases appetite via endocannabinoid receptors (CB1r); appetite stimulation is documented in animals [9, 10] and in healthy human [11, 12] and acquired immunodeficiency syndrome (AIDS) populations [13, 14]. Other potential benefits of THC therapy have been overlooked. Chemosensory alterations are common and distressing among advanced cancer patients, and these contribute to decreased food intake and enjoyment and diminished QOL [15, 16]. Patients frequently report loss of food ideation and desire to eat [17, 18]. THC may increase food intake by stimulating the orosensory reward pathway, increasing motivation to eat energy dense foods and enhancing food enjoyment [10, 19]. CB1r are located in reward-related areas of the brain [20] and in the olfactory epithelium and bulb [20, 21] and are involved in peripheral odor processing [21] and potentially taste function [22]. We hypothesized that THC may favorably alter chemosensory perception. We therefore undertook a randomized placebo-controlled trial to determine THC's therapeutic potential to improve food-intake behavior for cancer patients with self-reported chemosensory alterations. Our approach included a panel of patient-reported outcomes related to food-intake behavior including perceived chemosensory alterations, macronutrient preference, caloric intake, appetite, and QOL. Safety and tolerability were also assessed.
patients and methods
This two-centre (Cross Cancer Institute, Edmonton, Alberta and Jewish General Hospital, Montreal, Quebec), phase II, randomized, double-blind, placebo-controlled 22-day pilot study (NCT00316563) was approved by Health Canada and the Research Ethics Boards of the Alberta Cancer Board, University of Alberta, and McGill University.
eligibility criteria
Adult patients with advanced cancer (defined as locally recurrent, locally advanced, or metastatic) of any site except brain who had a score ≥2 (out of 16) on a scored Taste and Smell Survey [23] were included (Table 1, see below). The survey identifies and quantifies chemosensory alterations; a score of ≥2 is associated with decreased caloric and protein intake and poorer QOL compared with no chemosensory alterations (T. Brisbois Clarkson, I. De Kock, S. Watanabe, V. Baracos, W. Wismer, unpublished data). Included patients had decreased food intake for ≥2 weeks (reported by subject or physician) and a physician-assessed life expectancy of >2 months. All patients spoke English and provided informed consent. Use of chemotherapy and radiation therapy was permitted during the trial, provided that no therapy-related adverse events (AEs) ensued. Eligibility of patients was determined by review of patient charts and verified by study physicians. Once approved, patients were enrolled by a research assistant.
exclusion criteria
Exclusion criteria included receiving enteral or parenteral nutrition; allergies or sensitivity to THC and/or sesame seed oil; history of substance abuse (determined by review of patients' medical records, alcohol abuse was often also assessed by CAGE questionnaire [24]) or psychotic episodes (e.g. diagnosis of schizophrenia or psychosis); mechanical obstruction of alimentary tract, mouth, or nose; radiation therapy to the head/neck; primary brain tumor; nausea score >5 on 11-point scale (0 = no nausea, 10 = worst possible nausea); medical conditions affecting chemosensory function (i.e. infection of the mouth or nasal cavity, active sinusitis, hay fever), history of tachyarrhythmias, angina pectoris, or uncontrolled hypertension; liver impairment determined by Child-Pugh score ≥10; use of marijuana within 30 days before start of trial.
Patients on treatments with the specific intention of increasing appetite or anabolism were excluded (e.g. dexamethasone 4-10 mg b.i.d., megestrol acetate, cannabinoids, oxandrolone). Low-dose corticosteroids for indications other than appetite were allowed provided that doses remained constant for the duration of the trial. Patients were screened for oral candidiasis; if present, patients were eligible to participate once the infection was successfully treated.
random treatment assignment, blinding, and intervention
After baseline assessments, patients were randomly assigned by a third party pharmacist in a double-blinded manner to receive either THC (Marinol®; dronabinol 2.5 mg capsules, Solvay Pharma Inc.) or placebo according to a third-party computer-generated randomization scheme. Patients started on THC 2.5 mg or placebo once daily for the first 3 days (before bedtime for first 2 days and before supper on third day). The dose was increased to THC 2.5 mg or placebo twice daily (1 capsule before lunch and supper) on the fourth day (Figure 1). Patients had the option to increase their drug dose to a maximum of 20 mg/day [25].
outcome measures
Patients completed assessments at baseline (day 0) and after 18 days of treatment with 1 day latitude in posttreatment assessments (Figure 1). All assessments used patient-reported outcomes to capture and describe changes experienced and valued by the patients [26—28].
The Taste and Smell Survey [23] was used to identify and quantify chemosensory alterations 'since study treatment'. This tool has a maximum chemosensory complaint score of 16 (Table 1) [16]. Alterations in taste and smell emerging during study treatment could be pleasant or unpleasant. For example, a loss of sensation resulting in food being perceived as tasteless would be unpleasant, whereas the loss of a heightened and distressing sensitivity to tastes or odors could be perceived as a benefit. We therefore added the qualifying statements 'yes, it's better', 'yes, it's worse', and 'no, it's the same' to questions 1—4; 'chemosensory improvement' was calculated by tallying the 'yes, it's better' responses to these questions. Responses to questions 8—12 were further qualified as to whether change was 'pleasant or unpleasant'; responses of 'pleasantly stronger' comprised the 'chemosensory enhancement' outcome (maximum score of 5). The survey also includes open-ended questions for subjects to elaborate on the specific nature of chemosensory change(s) and their impact on QOL.
The 100 mm Satiety Labeled Intensity Magnitude (SLIM) scale [29] was completed 10—15 min before each meal for 1 day pretreatment and following 18 days of treatment for an assessment of appetite. This scale is anchored with greatest imaginable fullness = 0 and greatest imaginable hunger = 100 (neither hungry nor full = 50). The Macronutrient Preference Checklist (MPC) [30] was completed with the SLIM to assess concurrent macronutrient preferences. The MPC is scored on the number of food items selected (0—8) in each of four macronutrient categories (high protein, high fat, high carbohydrate, low energy). SLIM and MPC premeal scores were averaged for an overall day score. A 3-day dietary record [31] was used to estimate total calories and macronutrient intake (Food Processor II Nutrient Analysis Program; Esha Research, Salem, OR).
QOL was assessed with the Functional Assessment of Anorexia/Cachexia Therapy (FAACT) questionnaire [32]. The 11-point Edmonton Symptom Assessment System [33] was used to assess nausea. A Side Effect Survey [34] documented the tolerability of the study drug. Interviews were conducted to determine patients' treatment-related changes in food preferences and chemosensory alterations.
patient characteristics
Patients were recruited from either the palliative home care program or outpatient clinics at local cancer clinics in Edmonton, Alberta, and Montreal, Quebec, Canada, over 2.5 years (2006—2008). There were no differences for any outcomes between study sites (P < 0.05). Patient characteristics (Table 2) and dropout rates (Figure 2) were similar for THC and placebo groups. Of factors that could affect chemosensory perception, 33% of patients were receiving chemotherapy at the time of data collection (although all patients had previously received multiple rounds of various therapies). Similar proportions of participants in each treatment arm reported wearing dentures, experiencing previous mouth or gum infections, and previous taste or smell problems and were previous or current smokers. No patients were taking concurrent appetite stimulants. All participants were treated with various prior chemotherapies (Table 2); however, their effect on chemosensation is impossible to deconvolute in a sample of this size with diverse cancer diagnoses.
statistical analyses
Statistical analyses were performed on a per protocol analysis basis [35] using SAS (Statistical Analysis System for Windows, Cary, NC) [36]. Total and subdomain scores of the Taste and Smell Survey were treated as the primary outcomes, with secondary outcomes of appetite, macronutrient intake, nausea, and QOL. In this exploratory study, the nature of treatment effects and potential effect size(s) were unknown and for that reason, 10 participants per arm were assessed. Descriptive statistics were used to describe prevalence, nature, and severity of chemosensory alterations. Chi-square and Fisher's exact tests were used to evaluate patient characteristics, yes/no responses, treatment side effects, and AEs. Time series analysis of variance [37] with baseline assessments as covariates were significant [38] and were used to assess differences in chemosensory complaint scores, caloric intake, appetite, macronutrient preferences, QOL, and nausea between- and within-treatment groups as these data met assumptions of normality. Pairwise differences of least squares means (pdiff) were used for post hoc comparisons.
results
In the THC group, eight patients followed the dosing protocol (i.e. 2.5 mg b.i.d.) and three patients increased to 2.5 mg t.i.d. by taking an additional 2.5 mg before supper. In the placebo group, seven patients followed the dosing protocol and three patients increased their dose to 3 capsules/day.
taste and smell perception — results of the scored questions
Total chemosensory complaint scores decreased with THC treatment compared with baseline but were not different from placebo (Table 3). Chemosensory enhancement (frequency of response 'pleasantly stronger' on questions 8—12; Table 1) was increased from baseline with THC treatment compared with placebo (P = 0.018) (Table 3). Chemosensory improvement (frequency of response 'yes, it's better' on questions 1—4; Table 1) was twice as frequent with THC treatment (36%) compared with placebo (15%) (P = 0.026).
taste and smell perception — responses to open-ended questions and interview
Responses to open-ended questions and interview were concordant with the outcomes described above. The majority (73%) of THC-treated patients reported an increased overall appreciation of food compared with patients receiving placebo (30%) and more often stated that study medication 'made food taste better' (55%) compared with placebo (10%) (P = 0.04). Half of the patients who reported odors to be unpleasant at baseline no longer found odors offensive with THC treatment (P = 0.083). The majority of THC-treated patients (73%) indicated a renewed ability to discriminate tastes, flavors, and food odors. By contrast, 80% of patients in the placebo group reported their taste and smell function to be the 'same as before' (60%) or 'worse' (20%) compared with baseline.
appetite
For the THC group, SLIM appetite scores increased relative to baseline and placebo (Table 3). The majority of THC-treated patients (64%) had increased appetite, three patients (27%) showed no change and one patient's data were incomplete. No THC-treated patients showed a decrease in appetite. By contrast, the majority of patients receiving placebo had either decreased appetite (50%) or showed no change (20%).
food preferences and caloric intake
Compared with placebo, THC-treated patients increased their protein intake as a proportion of total energy. There was a trend for THC-treated patients to express preference for high protein foods on the MPC (Table 3). During interview, patients in the THC group (55%) commonly reported savory foods (e.g. hamburgers, chicken, fish, baked beans, and mushrooms) to 'taste better' and to be 'more appealing' since the study treatment. No patients in the placebo group reported an increased liking of meats.
Total caloric intake (Table 3) or intake per kilogram body weight (P = 0.557) did not differ between-treatment groups. Relative to baseline, 73% of THC-treated patients increased their caloric intake (range 100—775 kcal/day) compared with 50% patients in the placebo group (100—965 kcal/day).
quality of life
FAACT global QOL scores improved similarly for both THC and placebo groups (Table 3). The FAACT subdomain of anorexia—cachexia-related nutritional well-being improved in the THC group but was not different from placebo.
side effects and AEs
Nausea scores were unaffected by THC treatment (P = 0.532). Quality of sleep and relaxation were more frequently reported to be 'pleasant' by THC-treated patients compared with placebo on the Side Effect Survey (Table 3). There were no other differences in survey responses between-treatment groups (P > 0.05; Table 4).
THC was well tolerated. No differences were reported during the trial or within the 30-day follow-up period between THC and placebo groups for the number of AEs or serious AE (SAE) (P = 0.622 and P = 0.244, respectively). Most AEs were unrelated to THC therapy, six were unclear, and four were possibly related (Table 5). Most SAE were also unrelated to THC therapy, four were unclear and one was possibly related to treatment (irregular heart beat).
discussion
main findings
Our pilot study demonstrates that THC, compared with placebo, improved and enhanced chemosensory perception, altered macronutrient preference, appeal of savory foods, appetite, relaxation, and quality of sleep for advanced cancer patients with chemosensory alterations.
We opted for self-reported chemosensory perception as the most relevant predictor of food preference and enjoyment in lieu of objective clinical measures (i.e. millimolar concentration thresholds for detection of individual tastants and odorants). Along with Bartoshuk, we speculate that taste and smell alterations are not merely quantifiable physiological changes but also involve the loss of food enjoyment [18]. Bernhardson et al. [39] and Steinbach et al. [40] have recently used self-reported chemosensory perception to comprehensively describe taste and smell changes experienced by cancer patients. Clinical measures of chemosensation cannot capture dimensions such as flavor, food enjoyment, or impact on patient's food-intake behavior. The millimolar concentration threshold for perception of salt dissolved in water is not a surrogate for the complex perceptions that contribute to the ingestion and enjoyment of food.
Our findings parallel earlier results of enhanced sensory perception and improved food enjoyment among healthy marijuana users [41, 42]. THC also reduces the unpleasantness of a bitter taste solution in animals [19]. Our THC-treated patients reported odors and the taste of meat to be less offensive, possibly contributing to increased calories ingested as protein. These results suggest that THC improved chemosensory perception through reward systems [42].
Our findings are important as there is no accepted treatment for chemosensory alterations experienced by cancer patients [15, 16]. THC treatment may hold multiple clinical benefits for cancer patients, beyond its indication as a treatment for nausea and its effects on appetite. THC may palliate an array of symptoms: chemosensory alterations, food enjoyment, pain, depression, anxiety, poor quality of sleep, and inflammation [43]. Food enjoyment is a component of QOL, and while food enjoyment is intangible in terms of tools for its quantification, we nonetheless suggest that THC may well contribute to the overall enjoyment of food in cancer patients. Improved sleep may be due to the presence of CB1r in the basal forebrain [20] or related to increased relaxation noted in various populations [7, 44]. Here, improved quality of sleep and relaxation may have increased appetite and improved chemosensory perception, encouraging a positive outlook on food [45].
prior and future clinical trials of THC
THC stimulates appetite in healthy volunteers [11, 12] and AIDS patients [13, 14], but its ability to do so in cancer patients has not been consistently reported. THC increased appetite for 34%—72% of cancer patients with doses ranging from 5 to 45 mg/day [3—8, 44]. Nelson et al. [4] showed promising THC appetite stimulation in cancer and 6 of 18 patients opted to remain on THC treatment for improved appetite and food intake [46]. However, this study was criticized for the lack of a control group. Jatoi et al. [6] reported THC to stimulate appetite in 50% of patients. Strasser et al. [8] noted no differences between THC or THC + cannabidiol and placebo for appetite or QOL. In our study, 64% of THC-treated patients showed improved SLIM appetite scores and this assessment was not susceptible to placebo effect. The verbally anchored SLIM scale shows greater sensitivity and specificity than unlabelled visual analog scales [29], which had been used in other studies [8].
Jatoi et al. [6] and Strasser et al. [8] selected 5 mg THC daily to minimize side effects; however, Nelson et al. [4] showed more promising results with 7.5 mg THC daily. AIDS patients tolerated doses as high as 40 mg THC daily well [13]. We started patients at a low dose to build-up tolerance and minimize negative psychoactive effects [25] and allowed patients to titrate their dose upward and this was well tolerated. We noted dropouts and withdrawn consents due to changes in health status unrelated to study treatment. The dropout rate was not uncommon for an advanced cancer population [6]. Clinical trials in advanced cancer have the added complexity of comorbidities and expected death. The exclusion of data that are confounded by poor prognosis is critical for interpretable results, which may be a criticism of previous work [8].
One consideration for future trials will be the situation of therapy for taste and smell alterations within the disease trajectory, as these problems have been characterized both during first-line therapy in treatment naïve patients [47] as well as in advanced stage patients with progressive disease who have received multiple lines and cycles of treatment, as here. These may be related or distinct clinical entities.
Our sample size and study duration within a placebo-controlled design showed statistical significance for several outcomes responsive to THC: chemosensory perception, appetite, relaxation, and quality of sleep. However, our results require verification in larger trials. Our goal was to conduct a proof of principle study to provide direction for future trials. The data here would assist in the development of larger phase II trials by facilitating sample size calculations given our preliminary indications of effect sizes and variance (e.g. total chemosensory complaint scores as the primary outcome would require ∼50—60 patients in each treatment group based on the variance determined here). Our results offer a reasonable starting point from which future studies may investigate the use of THC in cancer anorexia where chemosensory alterations are present.
For the design of future trials, it seems important to (i) include a placebo group as outcomes may appear more favorable when compared with drug alone; (ii) include patient-reported assessments to capture relevant aspects of food-intake behavior, such as chemosensory and food preference changes; and (iii) power studies around differentiable outcomes, such as chemosensory complaint scores. As absorption of oral THC varies greatly between individuals [25, 48], and given the controversy surrounding the appropriate dose in cancer, future trials may allow patients to titrate their dose or compare different dose levels. Inevitably, questions are raised about the ability to blind THC treatment based on its well-known psychoactive characteristics; however, the timed administration of low doses and the lack of differences in AEs between-treatment groups suggest that this problem was likely minimal. THC merits further investigation as a therapy for patients who suffer from chemosensory alterations and loss of food enjoyment.
Source, Graphs and Figures: Delta-9-tetrahydrocannabinol may palliate altered chemosensory perception in cancer patients: results of a randomized, double-blind, placebo-controlled pilot trial
