Test Material
Relora® (NP 33–39) is a proprietary blend of a patented extract of the bark of Magnolia officinalis Rehder & Wilson [Magnoliaceae] (US Patent Nos. 6,582,735 and 6,814,987) and an extract of the bark of Phellodendron amurense Rupr. [Rutaceae]. The product is standardized to not less than 1.5% honokiol and 0.1% berberine. Subjects ingested 250 mg in dark red opaque capsules three times daily or a placebo that was identical in size, shape and color.
Study Design
In a prospective, two-group, parallel, randomized, double-blind, placebo-controlled trial design, healthy adult females with anxiety who reported that they typically eat more in response to stressful situations were enrolled and followed for six weeks. Participants were split into two groups of 20 using block code randomization (coded labels provided by the manufacturer) and then administered either Relora or a matched placebo for six weeks. The study was approved by the Integreview Ethical Review Board, Austin, TX.
Recruitment advertisements were placed in English and Spanish in local newspapers in the Miami, FL, area. The advertisements read, "Is stress making you eat? You may qualify to participate in a research study of a nutritional supplement for people who eat more during stressful situations. To qualify you must not be taking medications for depression or anxiety, be overweight but in good general health, be 20 to 50 years of age, and be a premenopausal female." The subjects for the study were pre-screened by phone and potential candidates were called in for a screening and baseline evaluation (Visit 1) after providing written informed consent. Acceptable subjects were enrolled (Visit 2), and then randomized with equal probability to receive either Relora or placebo. Efficacy and safety evaluations were performed during visits at treatment weeks 3 and 6 (Visits 3 and 4).
To be included in the study, subjects had to be healthy overweight females (BMI 25 – 39.9 kg/m2), be premenopausal (aged 25 to 50) and to self report increased eating in response to stress. Subjects were excluded if they had personal history of heart disease, uncontrolled high blood pressure, renal or hepatic impairment/disease, Type I or II diabetes, psychiatric disorders, cancer (except basal cell carcinoma of > 5 years ago), use of any monoamine oxidase inhibitor medication, sleep disorders, glaucoma, gastric ulcer or reflux disease or a seizure disorder, unstable thyroid disease, pregnancy, lactation, or any medical condition deemed exclusionary by the medical staff. Scores positive for Binge Eating Disorder utilizing the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria for diagnosis (score > 7), or scores positive for clinical depression via the Center for Epidemiologic Studies Depression Scale (CES-D) questionnaire (scores >16) or score below the norm for women 30 – 39 years of age on the State-Trait Anxiety Inventory (STAI) (< 34; per investigator discretion) or if the subject scored positively for severe anxiety they were also excluded from this study. Psychiatric disorders were determined by medical history, rechecked by the Principal Investigator (a medical doctor), and by means of questionnaires given by trained professionals. Subjects were also excluded if they were on monoamine oxidase inhibitors (MAO-I), anxiolytics, psychotropics including SSRI's, daily use of OTC medications or were currently taking stimulant medications (prescription, self-prescribed or OTC). In addition, recent use of weight control agents, unstable body weight, use of any form of a steroid medication, use of any type of dietary supplement were also excluded. Women who were pregnant, lactating, planning to become pregnant during the study or not using an acceptable form of contraceptive device or if a subject was ever diagnosed with Post-traumatic stress disorder (PTSD), they were excluded from this study. Study subjects were not allowed to take supplements that are, or were purported, to affect mood states. They were also not allowed to start any new supplements. If a subject was already taking a multivitamin/mineral supplement, they were allowed to continue doing so. To ensure compliance, a dichotomous questionnaire was used by the coordinator at each visit.
Analyses
Blood was drawn for analysis of biochemical markers for safety parameters at the beginning and end of the study. A 3-day (2 working days and 1 non-working day) food diary was recorded by the subjects at baseline and post-treatment. At the beginning, mid-point and end of the study, participants were weighed and tested using psychometric questionnaires. The psychometric questionnaires used in this study included: Spielberger STATE-TRAIT (20-item questionnaires by which subjects describe transient or general anxiety, on a 4-level scale); Visual Analog Scale – Sleep Quality (VAS-SQ; 10 cm scale, self-reported 5 point Likert scale, 0 = no sleep, 4 = very restful sleep); Sleep Latency Questionnaire – asked each time the VAS-SQ was conducted; subjective notion of how long, on average it takes the person to fall asleep, in minutes); and subjects were also given 3-day saliva collection kits (to be analyzed for salivary amylase and salivary cortisol) at the screening and mid-study visits (with samples taken at three different times (upon waking, 30 minutes later, and in the evening) for three consecutive days during each collection cycle), to be collected at the randomization and end-of-study visits, respectively.
Participants were also asked to follow their typical diets and exercise routines. Physical exercise quantity was quantified with the Framingham Physical Activity Index.
Statistical Plan
The primary efficacy analyses were conducted on an Intent-to-Treat basis, with missing observations imputed by the "last observation carried forward" (LOCF) method. Safety analyses were conducted on all subjects who received study product. Unpaired t-tests were used to compare changes over time between the two groups, except for variables that were substantially non-normally distributed, in which case the Mann-Whitney U-Test was employed. Statistical analysis was conduced using Excel 2002, SPSS version 13.0 and "R" version 2.1.1.
The number of participants for this study was pre-determined as a convenience sample. It was noted that a sample size of 20 subjects in each of two groups could provide 80% probability to produce a significant result (P < 0.05) in an unpaired comparison, when the mean between-group difference in any endpoint is 0.9 as large as the within-group standard deviation for that endpoint.