Subjects
We conducted this pilot study at The Preston Robert Tisch Brain Tumor Center at Duke. All subjects were newly diagnosed high-grade glioma patients undergoing standard concurrent chemotherapy and radiation therapy with radiotherapy and daily oral temozolomide (TMZ). Inclusion criteria for this study were: 1) written informed consent before beginning specific protocol procedures, 2) histologically proven high-grade glioma, 3) planned treatment with concurrent radiotherapy and daily oral temozolomide, 4) ≥ 18 years of age, 5) must be able to ambulate unassisted for 6 minutes safely, 6) hematocrit ≥ 29%, hemoglobin ≥ 9, ANC ≥ 1,500 cells/µl, platelets ≥ 100,000 cells/µl, and 7) serum creatinine < 1.5 times upper limit of normal, serum SGOT < 2.5 times upper limit of normal and bilirubin < 2.0 times upper limit of normal. Exclusion criteria for this study were: 1) prior therapy with naltrexone or naloxone, 2) co-medication that may interfere with study results; e.g opioids, 3) known hypersensitivity to any component of naltrexone, and 4) pregnant (positive pregnancy test) or lactating.
Study Design
This study was a placebo-controlled, randomized clinical trial comparing low-dose naltrexone and placebo and its effects on quality of life and fatigue. This study was approved by the Duke University Institutional Review Board and assigned number was Pro00027661 after receiving ethical approval from this designee. Patients were presented with the study and were provided with written informed consent. After obtaining written informed consent, we randomized patients to receive either placebo or LDN dosed at 4.5 mg orally to be taken every evening before going to bed. Using a permuted block algorithm, randomization was stratified by the inclusion of bevacizumab in the concurrent chemotherapy and radiation therapy treatment plan. We performed quality of life assessments using standardized, validated patient-reported outcome questionnaires at the following time points: 1. Baseline (before concurrent chemotherapy and radiation therapy), 2. After concurrent chemotherapy and radiation therapy (approximately eight weeks from the initial assessment), 3. Two months after standard concurrent chemotherapy and radiation therapy (approximately sixteen weeks after initial assessment), and 4. Four months after standard concurrent chemotherapy and radiation therapy (approximately 24 weeks after initial assessment). Treatment with LDN or placebo began on the first day of concurrent chemotherapy and radiation therapy and continued for a total of 16 weeks from the initial assessment. Duke University Compounding Pharmacy compounded LDN and placebo. Standardized dose of LDN was 4.5 mg, and patients administered LDN or placebo orally approximately 1 hour before bedtime.
Randomization
Subject randomization was stratified by the inclusion of bevacizumab in the concurrent chemotherapy and radiation therapy treatment plan. For each stratum, a randomization list was generated by the statistical team using a permuted block algorithm (block size = 4). This randomization list was provided to the investigational chemotherapy service so they could randomize patients to a treatment when the clinical research coordinator sent a fax with required information requesting patient randomization. Per study standard operating procedures, the statistical team reviewed the ICS logbook every three months to oversee the randomization process. The principal investigator clinical research team, and clinical team were blinded to treatment assignment throughout the duration of the study; only the statistical team and investigational chemotherapy service had access to treatment assignment. Subject and subject’s caregivers were blinded to treatment assignment.
Assessments
We evaluated the patient-reported quality of life with the Functional Assessment of Cancer Therapy-Brain (FACT-BR) scale. The FACTBR (version 4) contains subscales for physical (7-items), functional (7-items), emotional (6-items), and social/family (7-items) well-being. Also, this instrument contains a 23-item brain cancer subscale (BCS), which assesses symptoms commonly reported by brain cancer patients. We measured patient-reported fatigue using the 13-item Fatigue Scale using the Functional Assessment of Cancer Therapy-Fatigue (FACIT-Fatigue) subscale, version 4 [25]. Cella and colleagues demonstrated that a (+/-) 3-point change in fatigue scores suggests a clinically important difference [26].
We monitored all adverse events for patients on either LDN or placebo. Given that routine testing during standard of care therapy with concurrent TMZ and radiation therapy includes comprehensive metabolic with liver function panel and complete blood cell counts with differential, we did follow these laboratory values. During the protocol, all patients were required to have every other week monitoring of liver enzymes and liver function through liver function panels. For this protocol, we graded each adverse event per the guidelines in Common Terminology Criteria for Adverse Events (CTCAE) version 4.0.
Statistical Methods
Though this study is comparative, the goal of this randomized double-blinded pilot study is to determine whether LDN is worthy of further investigation, and not to make definitive statements about its effectiveness relative to placebo. Sample size determination focused on treatment differences in QOL at week 16 as measured by FACT-Br. Given the universal importance of a half standard deviation change or difference as previously reported for health-related quality of life measures, this study was designed to detect a half standard deviation difference in QOL scores measured at week 16 [27]. Being a phase II study, the sample size was constrained at the expense of either an increased false negative or false positive rate following recommending recommendations of by Ratain and Sargent [28]. This study, as originally designed, assumed that 10% of the original goal of 72 patients would withdraw from study participation before week 16. However, a preliminary look at this assumption by the data management and statistical teams confirmed a withdrawal rate of approximately 40% equally distributed between treatment arms. Hence, the sample size goal for this study was increased to 110 randomized patients to account for this higher rate of drop-out. With 55 randomized patients per arm in this screening study, we anticipated that approximately 33 patients per arm would provide a week 16 assessment. With 33 patients per arm, the power of a 1-tailed two-sample t-test conducted at the 0.2 level of significance to detect a half standard deviation difference in QOL scores measured at week 16 is 88% (SAS 9.2, Cary, NC).
Descriptive statistics (e.g. means and standard deviations) summarized within-group changes from baseline in each FACT-Br and FACIT-Fatigue subscale at each follow-up assessment. Analysis of covariance will be used to compare groups with respect to changes between baseline and 2 months post concurrent radiation therapy and chemotherapy in each of these subscales, with the baseline measure included as a covariate.