This trial was performed at Alluri Sitarama Raju Academy of Medical Sciences (ASRAM), Eluru, Andhra Pradesh, India from July 2006 to October 2006 (clinical trial registration number: ISRCTN05212803). The study protocol was evaluated and approved by the ASRAM Institutional Review Board. An overview of the clinical study is provided in Figure 1. Briefly, 236 patients out of 823 attending the orthopaedic Outpatients Department of the ASRAM Hospital were selected, based on the signs, symptoms and radiological changes consistent with OA in the first phase of the screening procedure. A total of 75 patients suffering for more than 3 months with medial tibiofemoral OA were selected using inclusion/exclusion criteria summarized in Table 1. All patients signed the Institutional Review Board approved consent form. Patients were otherwise healthy, were aged 40 years or older, and had a diagnosis of OA, fulfilling the American College of Rheumatology classification criteria 4. After recruitment, the patients were randomly distributed into three groups; demographic data and baseline characteristics are summarized in Table 2.

Before study enrollment, patients were required to be taking an NSAID at prescription strength for at least 30 days or acetaminophen 1,200 to 4,000 mg/day on a regular basis (at least 25 of the preceding 30 days) with a history of therapeutic benefit. Eligibility required patients to meet specific flare criteria upon medication washout. At screening, patients had to demonstrate a visual analog scale (VAS) score between 40 and 70 mm during the most painful knee movement, and Lequesne's Functional Index (LFI) score greater than 7 points after 7-day withdrawal of usual medication.

A total of 75 selected patients with symptoms of moderate to mild OA were recruited into the study. Each patient was randomly assigned to a treatment group using a randomization table generated using validated computer software (RANCODE; IDV, Gauting, Germany). Treatment allocation depended only on the time sequence in which patients entered the study, thus minimizing selection bias. The clinical trial pharmacist and statistician ensured that treatment codes remained confidential. The patients were distributed into three groups: placebo (n = 25); 30% AKBA enriched B. serrata extract (5-Loxin^®) low-dose group (100 mg/day), in which patients received 50 mg encapsulated 5-Loxin^® twice daily (n = 25); and 5-Loxin^® high-dose group (250 mg/day), in which patients received 125 mg encapsulated 5-Loxin^® twice daily (n = 25). Patients in the placebo group received two capsules of similar color, taste and appearance but filled only with rice bran.

Each patient completed a questionnaire, providing details regarding demographics, medical history and nutritional status, at the baseline evaluation and during the follow-up evaluations on days 7, 30, 60 and 90. At the baseline evaluation, and at each visit during the 90-day follow up period, all patients were assessed for pain scores and physical ability. Various parameters of serum biochemistry, haematology and urine analysis were carried out on each evaluation day. Serum samples were collected at all evaluation days for proinflammatory modulators. Knee joint synovial fluid was aseptically collected at baseline and at day 90 for evaluation of MMP-3 concentration. Safety was monitored by clinical and laboratory assessments conducted at study visits and patient-reported adverse experiences.

The investigators assessed the functional disability reported by the patients at baseline and on each follow-up visit (days 7, 30, 60 and 90). Questionnaire-based assessment of pain, stiffness and physical function were done using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) index 20, LFI 21 and VAS 22. The WOMAC index produces scores for three subscales: pain, stiffness and physical function. The pain, stiffness and function subscales of the WOMAC were converted to a 0 to 100 normalized units (NU) scale 23. The pain subscale was the average of the first five questions of WOMAC and measured using the NU scale from 0 mm ('no pain') to 100 mm ('extreme pain') for each question. The stiffness subscale was the average of questions 6 and 7, measured using the NU scale from 0 mm ('no stiffness') to 100 mm ('extreme stiffness') for each question. The physical function subscale was the average of questions 8 through 24 of the WOMAC and measured by NU scale from 0 mm ('no difficulty') to 100 mm ('extreme difficulty') for each question. Analyses of these end-points were based upon the time-weighted average change from baseline over 90 days.

For assessment of safety of 5-Loxin^®, several parameters were evaluated in serum, urine and whole blood of all patients at each visit of the study duration (Table 3). Serum biochemical parameters and haematological parameters were measured using the automated analyzer HumaStar 300 (Human, Wiesbaden, Germany) and the haematological counter Humacount (Human), respectively. The urine analysis was carried out by microscopy and by using UroColor™10 Dip Sticks (Standard Diagnostics, Kyonggi-do, Korea).

MMP-3 (R&D Systems, Minneapolis, USA) were quantitatively measured by ultrasensitive ELISA method. Assay procedures adhered to the protocol supplied by the manufacturers. Briefly, synovial fluid samples were incubated on capture antibody coated 96-well microplates. Specifically bound antigen was detected by appropriate biotinylated detection antibody and was probed with horseradish peroxidase enzyme. The specific immune reaction was detected by substrate solution and the colour development was read with the help of micro-plate reader (Bio-Rad, Hercules, CA, USA). A standard curve was generated by plotting the optical density at respective known concentration of MMP3. The sensitivity of MMP-3 detection ELISA kit is 9 pg/ml.

Patients were prescribed ibuprophen 400 mg tablets (maximum 400 mg thrice daily; total 1,200 mg) as rescue analgesia on days 7, 30 and 60, based on pain intensity reported to the study physician by the patient. However, the patients were instructed not to take medicine at least 3 days before each evaluation. No other OA interventions were allowed during the study period.

We performed detailed statistical analyses using SAS software to evaluate the efficacy of two doses of 5-Loxin^® in comparison with the placebo group in terms of improvement in pain and physical ability scores, and to assess biomolecular markers at baseline and days 7, 30, 60 and 90 of treatment. Pair-wise changes were examined by carrying out a least significant difference test for all possible pairs. The significance of the effects of the treatment groups was compared by using one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison tests. Results with P < 0.05 are considered statistically significant.

This is a three-arm (two doses of 5-Loxin^® and placebo), randomized, double-blind, placebo-controlled, single-centre trial conducted over 90 days. The trial's primary objective was to determine the effects of 5-Loxin^® on pain, physical function and joint stiffness. For power calculations, the estimates for variability and assumed mean changes for each treatment group were based on results from previous placebo-controlled studies of celecoxib, etoricoxib and rofecoxib conducted in patients with OA 24252627. We believe that an intervention that gives an average improvement of mean change + 1 standard deviation, rather than mean change only, will provide results of greater significance 28. Our trial is designed to have more than 80% power to detect a situation in which either active drug dosage yields an improvement to at least mean change + 0.9 standard deviation, under a conservative assumption, and we tested differences between groups in mean improvement using ANOVA (α = 0.05, two-sided). With 25 patients per group, we would have a 93% chance of observing at least one example of any side effect occurring in 10% or more of the patient population at a specific dosage.

Descriptive statistics comparing demographic variables, baseline disease characteristics and baseline outcome measures (that is, WOMAC pain, function and stiffness subscores) are provided in Table 2. Overall, the treatment groups receiving 5-Loxin^® low dose (100 mg/day, n = 25), 5-Loxin^® high dose (250 mg/day, n = 25) and placebo (n = 25), were similar with respect to sex, age, Body Mass Index and pain severity (Table 2). The patients were randomly distributed into three groups. Although there are some differences in baseline characteristics of gender, body mass index and WOMAC scores, those are statistically not significant.

We compared the scores between the treatment groups obtained at day 90. Both doses of 5-Loxin^® conferred clinically and statistically significant improvements in pain scores and physical ability scores in OA patients between baseline and day 90 (Table 4).

Tukey's multiple comparison test revealed statistically significant improvements by 48.83% (P < 0.001), 23.79% (P < 0.036) and 39.61% (P = 0.009) in VAS, LFI and WOMAC pain scores, respectively, in the low-dose (100 mg 5-Loxin^®) group versus the placebo group (Table 4). Improvements by 42.5% (P = 0.120) and 28.62% (P = 0.100) score in WOMAC stiffness and WOMAC functional ability, respectively were also achieved in the low-dose group (Table 4).

In comparison with the placebo group, the high-dose (250 mg 5-Loxin^®) group also exhibited statistically significant improvements in all parameters (Table 4). The high-dose group showed improvements by 65.94% (P < 0.001), 31.34% (P < 0.017), 52.05% (P < 0.001), 62.22% (P = 0.014) and 49.34% (P = 0.002) in VAS, LFI, WOMAC pain, WOMAC stiffness and WOMAC functional ability scores, respectively.

Student's t-test analyses revealed that MMP-3 concentration (P < 0.0001) in synovial fluids and VAS pain scores (P = 0.001) were significantly lower in the high-dose group than in the low-dose group. It is worth noting that both low-dose and high-dose treatment groups exhibited improvement in pain scores and physical ability scores as early as 7 days after the start of treatment, and these indices continued to improve throughout the 90 days of treatment (Figure 2). After 7 days, the low-dose and high-dose treatment groups exhibited 10.09% (P = 0.05) and 12.18% (P = 0.02) reductions in VAS, respectively, compared with the placebo group. In addition, WOMAC physical ability also improved by 14.38% (P < 0.01) after 7 days of treatment with high-dose 5-Loxin^® (Figure 2).

OA is a degenerative joint disorder; in molecular pathogenesis of OA, proteolytic enzymes such as MMPs are highly elevated in body fluids such as serum and synovial fluids, which cause potential damage in cartilage tissues 29. Therefore, in order to determine whether 5-Loxin^® treatment can normalize the MMP level, we evaluated the concentration of MMP-3 in synovial fluids collected from the patients. Figure 3 illustrates changes in MMP-3 concentration in synovial fluid samples collected from patients of all groups included in the study. Pair-wise comparisons indicated that at the end of the study both treatment groups exhibited highly significant reductions in MMP-3 in synovial fluid. Compared with the placebo group, the low-dose (100 mg) and high-dose (250 mg) 5-Loxin^® groups showed 31.37% (P = 0.002) and 46.4% (P < 0.001) reductions in MMP-3 concentration, respectively. Regular t-tests revealed that high-dose 5-Loxin^® treatment significantly reduced (P < 0.0001) synovial MMP-3 concentration when compared with the low-dose group (Table 4). Compared with baseline, the Wilcoxon sign-rank-sum test revealed that the low-dose and high-dose groups conferred 28.69% (P = 0.0013) and 46.33% (P < 0.0001) reductions in synovial fluid MMP-3 concentration at day 90. The MMP-3 level in the placebo group remained virtually unchanged at day 90 compared with baseline.

As a part of the safety evaluation, laboratory tests were performed to evaluate different biochemical parameters in serum and urine, and haematological parameters. The tested parameters in serum biochemistry, and haematological and urine analysis are summarized in Table 3. The significance of the differences between baseline and 90 days was tested by using repeated measures ANOVA. The F ratio is considered significant if P < 0.05. Although minor changes were observed in some of the parameters, they remained within the normal laboratory range. Statistical analyses of these parameters did not identify any statstically significant changes. Similarly, haematological and urinary parameters also exhibited no significant changes in the active treatment groups compared with placebo (data not shown). These findings further demonstrate the safety of 5-Loxin^® in humans.

During the course of the 90-day study period, some minor adverse events were noted: diarrhoea, nausea, abdominal pain, mild fever (up to 37.5°C [99.5°F]) and general weakness. The patients who reported these minor events were distributed evenly throughout the placebo and active treatment groups. The numbers of minor adverse events reported by the patients during the study are summarized in Table 5.

Five patients (one from the low-dose [100 mg 5-Loxin^®] group, and two each from placebo and high-dose [250 mg 5-Loxin^®] group) were excluded from the study because they were suffering from a nonfatal viral infection during the course of study.