Borage and Echium Seed Oils for Asthma

To evaluate the mechanisms by which borage seed oil and echium seed oil block mediator generation in asthma, we performed a cross-over trial design to reduce the numbers of patients needed for the study. Each treatment period was six weeks with a six-week washout period. For simplicity all power calculations are made assuming a single factor or predictor variable and a single result or outcome variable. We focused on changes in mediator generation by effector cells. Calculations below suggest that we have resolution to pick up clinically meaningful effects of potentially important explanatory variables with acceptable or high power.

1. Study Design The goals were to examine the mechanisms that underlie the effects of borage seed oil and echium seed oil on proinflammatory mediator production by effector leukocytes.

In the treatment arm, subjects consumed 4.0 g/day borage seed oil and 7.0 g/day echium seed oil (containing totals of ~1.6 g/day of GLA and ~0.9 g/day of SDA). The oils were packaged in capsules each of which contained 1 gram of oil. Therefore subjects will ingest 10 capsules a day (3 in the morning, 3 in the afternoon, and 4 in the evening) with meals. In the placebo arm, subjects took matching capsules containing corn oil; three 1 gram capsules in the morning and afternoon, and four in the evening, with meals.

The study began with a screening visit. All subjects had physician-diagnosed asthma. All performed a baseline set of lung function measurements. They entered a 2-week run-in period during which they kept a diary of peak flows, asthma symptoms, and beta agonist use. Subjects were instructed in peak flow technique and asked to perform three maneuvers each time. The maneuver was considered technically sufficient if the variation in the values obtained on the three attempts are with 10% of each other, and the highest value was recorded in a diary. Subjects whose diaries were > 80% complete entered the study and were randomly assigned to active treatment or placebo. Spirometry was repeated. Blood was taken for fatty acid composition of plasma and leukocytes, LT generation, and DNA for genotyping at the LTC4S locus. Subjects were provided with peak flow diaries, instructed in the appropriate performance or peak flow maneuvers, and asked to record morning and evening peak flow each day (reflecting the best of three efforts each time). After three weeks of treatment, diary cards will be collected and spirometry repeated. After six weeks of treatment, blood was drawn for safety monitoring, for measurements of fatty acids in plasma and leukocytes, and for the functional analyses. Spirometry was repeated. Subjects then entered a six-week washout period. Two weeks before the end of this period they again started to keep diary cards. At the conclusion of the 6-week washout period they entered the cross-over treatment phase (six weeks) which followed the same protocol as the first treatment period with assessment of fatty acid analyses, cellular function studies, safety monitoring, and spirometry. Pregnancy testing was performed before and after each treatment period. Dairy cards (with peak flow monitoring results) will be collected at each visit (3 week intervals during the placebo and active treatment arms of the study).

1. Statistical Analyses
2. Power Calculations - For simplicity all power calculations are made assuming a single factor or predictor variable and a single result or outcome variable. While a number of correlated outcomes will be investigated, we will focus on changes in mediator generation and levels of enzyme expression. Calculations below suggest that we have resolution to pick up clinically meaningful effects of potentially important explanatory variables with acceptable or high power.
3. Biochemical Endpoints - Power is based on paired t-tests comparing continuous measures of response between treatment and placebo. We assume 40 subjects will be available after allowing for a drop out rate of 20% from our original sample of 50 subjects. For the endpoint of leukotriene production by leukocytes, the power calculations suggested that we will have 80% power to detect a true within-subject difference of 0.8 of the within-subject SD between treatment and placebo with an α value of 0/05. On the basis of our previous studies, we assume a within-subject coefficient of variation (CV) of 13% for LTC4 generation. This will translate to an approximate detectable difference from 10.5% in these parameters.Using a paired two-sided t-test and prior mean(standard deviation) estimates for the placebo of .001464 ( .002268) and for the borage oil of .000289( .000709) , we would need 40subjects for 80-90% power at a 0.05 level of significance to detect a change in LTC4S transcript, protein, and enzymatic activity of 50% between placebo and borage arms.
4. Analysis Plan - Initially, we will present descriptive statistics (mean, standard deviation, range) for all endpoints to treatment and placebo groups. We will also present graphical plots of all measures across the study and for each group within each period. For the 2x2 crossover study, we will compare the treatment and control on each endpoint separately by employing a generalized mixed model approach with the predictors: group(sequence), patient nested within group as a random effect, period(dichotomous), treatment(dichotomous), and seasonality(dichotomous), with unstructured covariance. We will utilize mixed model diagnostics such as residual plots to assess goodness of fit.