Glucosamine sulphate was purchased from Rottapharm Ltd (Neptune, New Jersey, USA) and given as one 1500 mg powder in 295 ml of water. Glucose tolerance test solution, in Trutol Cola 75 g/bottle, was purchased from NERL Diagnostics (Fischer Scientific, East Providence, Rhode Island, USA).
Subjects fulfilling the American College of Rheumatology criteria for hand, hip or knee osteoarthritis, but otherwise apparently healthy were recruited over a period of several months from Tufts‐New England Medical Arthritis Treatment Center (Boston, Massachusetts, USA). Subjects were excluded if they had a body mass index (BMI) <20 or >40 kg/m2, known diabetes mellitus, fasting blood glucose level >110 mg/dl, kidney or liver disease, acute illness, previous myocardial infarction, uncontrolled inflammatory condition, malignancy, endocrine disorder, corticosteroid use, anaemia, or pregnancy. Informed consent was obtained with approval by the Human Institutional Review Boards, Tufts‐New England Medical Center and Bedford Veteran's Affairs Hospital (Bedford, Massachusetts, USA).
Of the 18 subjects who began the study, two withdrew after the first visit for reasons unrelated to the study protocol. Analyses were performed for the 11 women and 5 men who completed all study visits. Age (41–74 years, median 62 years), weight (42–132 kg, median 91 kg), BMI (22–40 kg/m2
,median 31 kg/m2
), area affected by osteoarthritis and previous use of glucosamine by subjects were as published previously.14
The study was performed in three visits, 1–2 weeks apart, and a fourth, several months later, with subjects fasted overnight and all morning drugs withheld. The doses at each visit were as follows: 1500 mg of glucosamine sulphate at visit 1; 75 g of glucose at visit 2; 1500 mg of glucosamine sulphate then 75 g of glucose given within 5 min at visit 3; and control without glucose or glucosamine sulphate at visit 4.
An intravenous catheter was inserted and kept clear with normal saline. Blood was obtained at time 0 and then at 15, 30, 45, 60, 90, 120, 150 and 180 min after ingestion. After clotting at room temperature for 30 min, samples were centrifuged at 1500 rpm at 4°C for 15 min and serum stored at −70°C until assay.
To measure glucose, 50 μl of serum was diluted 1:40 by addition of 1.95 ml of water for automated analysis at 35°C by a Metrohm‐Peak 817 Bioscan (Metrohm‐Peak, Inc., Houston, TX, USA) by pulsed amperometric detection after elution from an ion‐exchange Metrosep Carb 1 (250×4.5 mm) column, and measured against standards as described previously.14
Serum insulin was measured by radioimmunoassay (‘Coat‐a‐count' radioimmunoassay kit, Diagnostic Products Corporation, Los Angeles, California, USA) with an interassay coefficient of variation <10%.
Oral glucose tolerance profiles were classified as normal (2‐h glucose, <140 mg/dl); impaired (2‐h glucose, >140–<200 mg/dl); or diabetic (2‐h glucose, >200 mg/dl) by World Health Organization criteria. Areas under the curve (AUCs) for increase in glucose and insulin over baseline were calculated by trapezoidal rule, and means were tested for significance by paired t test. Mean glucose and insulin blood levels of all subjects were calculated for each time point.