There were totally 21 patients with hypercholesterolemia on treatment with a stable dose of simvastatin (10–40 mg once daily) and 13 patients with atorvastatin (10–20 mg daily) that were included in the study. The 11 males and 10 females in the simvastatin group had a mean age of 66 years, range 57–77 years. Fourteen patients had hypertension, nine type 2 diabetes and five had suffered from cardiac infarction or stroke (Table ). In the atorvastatin group there were nine men and four women with a mean age of 67 years, range 47–75 years. Eight patients had hypertension, five had diabetes, and four had earlier experienced heart infarction or stroke (Table ).
The study was approved by the Swedish Medical Products Agency and the Regional Ethical Board in Western Sweden.
The study was performed as a randomised, open, cross-over study, where the approved glucosamine containing product Artrox® was compared with an inactive control (a commercially available vitamin product named Vitamineral®). Patients were recruited from the primary health care centres in Mölnlycke and Landvetter in the Western region of Sweden. Eligible patients on simvastatin or atorvastatin treatment were identified by the routinely used data system at the primary health care centres. Patient characteristics are given in Table . Patients were invited by an introductory letter which included information about the study. The information leaflet was approved by the regional ethical board and it was sent home to eligible patients. Those who were willing to participate signed an informed consent and were included in the study. No record was made of those patients who did not respond to the information letter.
Exclusion criteria were; unstable angina pectoris, recent myocardial infarction (within a year), recent stroke (within a year), cardiac failure, HIV and dementia. There have not been shown any significant interactions involving glucosamine except case reports to warfarin treated patients. However, we excluded same drugs as in our earlier trials with St John´s Wort as our hypothesis was a stimulation of CYP3A4. Thus, treatment with birth control pills, warfarin, theofylline, cyclosporine, amitryptilin, nortriptilin, digoxin or sertraline was not allowed.
The study started with a four week run-in period. At the initial visit (week 0), the patients were informed about the study, inclusion and exclusion criteria were verified and a physical examination was performed. All blood samples were obtained in the fasting condition and analysed at the accredited central laboratory at Sahlgrenska University Hospital in Gothenburg (total s-cholesterol, s-HDL-cholesterol, s-LDL-cholesterol and s-triglycerides). The laboratory was unaware of patient treatment. Patients were then to return after 4 weeks (week 4) and another physical examination was performed and blood samples as described at week 0 above were collected. At this visit, the treating physician randomised (using sealed envelopes performed by a person not involved with patients), the order of treatment with active (Artrox®) and control (Vitamineral®). Thus, 50% started with active treatment and the other 50% started with control treatment and all were crossed- over to the other treatment modality after another 4 weeks (week 8). We considered that the treatment period o four weeks was adequate for possible glucosamine effects to vanish and thus no further wash out period was mandatory. Assessment of total s-cholesterol, s-HDL and s-LDL-cholesterol and s-triglycerides were then performed at the end of each treatment period (week 8 and 12). Compliance was verified by pill counting at each visit. The patients were on their regular dose of simvastatin or atorvastatin and the dose of simvastatin and atorvastatin was kept unchanged during the whole study period until all patients completed the study. The mean dose of simvastatin was 21 mg (range 10 mg – 40 mg) and of atorvastatin 14.6 mg (range 10–20 mg).
Glucosamine could have an effect on glucose metabolism as stated in the SPC of the product. In order to exclude negative effects on this variable, HbA1c was followed as a safety measure,
T- test for paired samples was used for statistical analyses and a p-value <0.05 was considered significant. The primary efficacy endpoint was the difference in LDL-cholesterol at week 8 and 12 between active treatment and control. Secondary endpoints were the corresponding difference in total cholesterol, HDL- cholesterol and in triglycerides. The study was powered to detect a difference of 0.48 mmol/L with a power of 81.2 % with a significance level of 0.05 (two sided) and a standard deviation within the groups of 0.52. Power calculations showed that participation of 20 subjects would give statistical significance. Randomization procedure was performed by a person not involved with the patients.