The study was conducted as a randomized, double-blind, placebo-controlled trial from September 2003 to April 2004 at the University of Alberta, Edmonton.
Subjects were recruited through media advertisements from Edmonton and the surrounding areas. Volunteers who responded to the advertisements were initially screened by telephone for inclusion and exclusion criteria. They were required to be in good general health, between 18 and 65 years of age and to have contracted at least 2 colds in the past year. To avoid the confounding effects of vaccination, subjects were excluded if they had been vaccinated against influenza in the previous 6 months. Subjects with medical conditions such as multiple sclerosis, tuberculosis, diabetes, cancer, lupus, HIV/AIDS, cardiovascular disease, hypertension, neurologic or psychiatric disease, and renal, pulmonary and hepatic abnormalities were excluded. Subjects taking medications such as immunosuppressive drugs, corticosteroids, warfarin, phenalzine, pentobarbital, haloperidol or cyclosporine were also excluded, as were pregnant or lactating women and heavy smokers. Because the potential impact of the North American ginseng extract in patients with these conditions has not been specifically examined in clinical studies, these exclusion factors were chosen on the basis of the results of a systematic review of the safety of Asian ginseng.17
Volunteers who qualified were asked to report to the University of Alberta for an information session. At these sessions the subjects completed a questionnaire pertaining to the history of past cold infections. The study protocol was also explained, consent forms were signed and a 2-month supply of either the ginseng extract or placebo was distributed along with the information sheet and daily cold assessment questionnaires. The standardized extract, for which batch-to-batch consistency is ensured in the production process, contains 80% poly-furanosyl-pyranosyl-saccharides and 10% protein formulated from the roots of North American ginseng (P. quinquefolium). The freeze-dried extract was encapsulated to contain 200 mg/capsule. The placebo was rice powder, which was encapsulated identically to the active treatment.
Subjects were randomly assigned to receive either the ginseng extract or placebo using an unrestricted randomization scheme generated using Excel 2003. EGA Biosciences (Edmonton, Alberta) administered the randomization scheme, providing investigators with numbered, opaque, sealed envelopes containing the treatment codes. Separate envelopes were to be opened only in case of medical emergency. The randomization codes were not broken until all of the data were analyzed.
Subjects were instructed to take 2 capsules per day for a period of 4 months following the onset of influenza season (November 2003). The study physician, G.P., notified the investigators when to start the treatment. Those with respiratory infections at the start of treatment were asked to delay beginning until absolute recovery.
Subjects were asked to take the capsules daily, in the mornings, after breakfast with a glass of water. They were instructed not to take any other cold medication unless advised by their family physicians. Subjects were contacted by e-mail or telephone each month to assure adherence to the study protocol. Compliance was also verified by weighing the returned bottles.
Subjects were asked to complete a daily log at about the same time every evening, to document the severity of their cold-related symptoms (sore throat, runny nose, sneeze, nasal congestion, malaise, fever, headache, hoarseness, earaches and cough) on a 4-point scale (0 = no symptom, 1 = mild symptom, 2 = moderate symptom, and 3 = severe symptom). The total symptom score3,4,18
was calculated by summing the daily scores for all symptoms. Subjects were told to contact the investigators and the study physician at the onset of a cold. A 2-day total symptom score greater than 14 (modified Jackson criteria)19
was considered to indicate a Jackson-verified cold; these were used in the analysis of number of colds. A daily total symptom score exceeding 4 was used in the analysis of symptom severity, number of days cold symptoms were reported, and duration of colds. This assessment was further verified by telephone or e-mail contact. The study physician also interviewed subjects, enquiring about the presence of symptoms that suggested secondary complications such as sinusitis, bronchitis, otitis media or pneumonia, and recommended family physician follow-up if necessary. To assess the tolerability of the treatments, subjects were asked to record any adverse side effects.
Subjects returned to the university at 2 months to receive additional medication and again at 4 months. At these visits they returned their daily assessment forms and unused medication. To ensure the integrity of randomization, identically numbered bottles (4 per subject), each containing 70 capsules of either the ginseng extract or placebo, were supplied by EGA Biosciences. These were distributed in numerical order. The numbers on the bottles were also used as the subjects' identification number. During the second visit the investigators as well the study subjects verified that the bottles returned were replaced with identically numbered bottles. As well, after completion of the intervention, subjects were asked whether they thought they had received the ginseng extract or placebo (“What do you think you received — ginseng extract or placebo?”).
All analyses were performed by a statistician under blinded conditions. A sample of about 123 in each group was found to be adequate. This was determined assuming the mean number of colds in the control group to be 2.5 per subject (primary efficacy variable), a treatment efficacy of 30%, a standard deviation of 2.0, a significance level of 5% and a power of 80%.
Our primary efficacy end point was the number of colds reported and Jackson-verified per subject. The secondary efficacy variables included severity of symptoms (total symptom score), number of days symptoms were experienced and duration of all colds during the 4-month intervention period.
The data were analyzed for all of the subjects who had been randomly assigned to a group, excluding those who provided only baseline information. An intention-to-treat analysis was performed. For those who discontinued the study, the last value available was carried forward. Those with incomplete daily logs were called and the scores obtained.
The 2 groups were compared with respect to baseline characteristics. The primary analysis was performed using an unpaired t test on the log-transformed data, with a Satterthwaite adjustment made to the degrees of freedom to allow for unequal variance. We used χ2 tests to compare the percentages of subjects contracting colds and receiving additional medication, and Fisher's exact 2-tailed probability test to compare the percentage of subjects reporting adverse events. Unpaired t tests were used to compare the 2 groups in terms of the number of colds, symptom scores and duration of symptoms.
The study was approved by the Human Ethics Committee of the University of Alberta.