RESULTS
A total of 102 patients with RA (93% women) were studied from February 2006 to April 2007; 79% of the women were housewives. The mean (SD) age of the patients was 48.6 (12.0) years. About two-thirds of the patients were seropositive. A majority of patients (57%) had a Steinbrocker Class 2 functional status.
   The current prevalence of CAM use in our study population was 39%, the all-time prevalence being 82%. A total of 215 courses were reported with 77 courses being used at present (mean of 2.1 CAM courses per patient). As compared to non-users, CAM users had a significantly higher mean age (50.1 years v. 41.2 years; p=0.016). CAM users had a longer duration of disease compared with non-users (12.27 years v. 9.16 years; p=0.059), which was statistically not significant. There was no significant association between CAM use and gender, marital status, occupation or socioeconomic status of the patient. The commonly used CAMs by this study group in decreasing order of prevalence were ayurveda, homoeopathy, yoga asana, pranayama, massage, Unani and acupressure (Fig. 1). However, among the 77 CAM therapies being currently used, yoga asana was the most frequently employed followed by pranayama, massage, homoeopathy, ayurveda and acupressure. The mean duration (SD; median) in months for which different CAM therapies were used were as follows: yoga asana 37 (58.5; 12); pranayama 27.5 (59.5; 8) followed by homeopathy 16.8 (30.6; 6) and ayurveda 16.2 (23.4; 6). Yoga asana and pranayama were the most frequently used currently and tended to be adhered to for a longer duration compared with ayurvedic and homoeopathic therapies (p=0.01).
   The most common reason for starting CAM use was control of pain. Up to 69% of CAM users reported initiation of CAM solely for this reason. The other reasons cited for initiating CAM therapy were pressure from family and contacts (11.9%), dissatisfaction with allopathic medicines (5.9%) and adverse effects of allopathic medicines (5.9%). It was observed that therapies such as yoga asana (84%) and pranayama (88%) were more often started after the institution of allopathic treatment. On the other hand, ayurveda (73%) and homoeopathy (72%) were more often started after diagnosis and before the initiation of disease-modifying drugs. Among CAM therapies, 83.7% of the yoga asana and 92.3% of the pranayama courses were being continued at the time of the study compared with only 6.7% of ayurveda and 13.9% of homoeopathy courses.
   A majority of the CAM courses (78.1%) were begun on the advice of friends/neighbours/relatives. A trained CAM provider was identified for influencing the commencement of CAM therapies in only 21% of courses. The media influenced the adoption of 16% of CAM courses. Yoga asana and pranayama were often (52%) adopted based on media information. On the other hand, 75% of ayurveda and homoeopathy courses were taken up on the advice of close associates. At the time of the study, 138 CAM courses were reported to have been discontinued by the

* Questions used for the variables of CAM disclosure were adapted with permission from Rao JK, Mihaliak K, Kroenke K, Bradley J, Tierney WM, Weinberger M. Use of complementary therapies for arthritis among patients of rheumatologists. Ann Intern Med 1999;131:409–16.

patients with lack of clinical benefit cited as the major reason for termination of 78% followed by adverse effects (10%). No adverse effects were ever reported with the use of yoga asana or pranayama.
   The mean (SD; median) expenditure on CAM was calculated as Rs 793 (1761; 100) per month per course (movement-based therapies such as yoga asana Rs 38 [170; 0], pranayama Rs 0, ayurveda Rs 782 [1413; 400] and homoeopathy Rs 707 [1162; 200]).
   Of the 84 patients who had ever used CAM, only 23 (27.3%) had revealed its use to the physician (Table I). The reasons for this and the reactions of the physician to the fact that they were using CAM are also summarized in Table I. A vast majority (87%) of those who did not reveal their CAM use to the physician said that the physician did not enquire about CAM use. Only 5% of patients did not inform their physician due to fear of disapproval. No association was seen between the physician–patient discussion of CAM use with female gender, college education or disease characteristics such as the functional status score, disability index and mean duration of disease.

DISCUSSION
Our study reveals that CAM use by RA patients is very high. As many as 82% of the patients reported having ever used CAM; the current use of CAM in our study group was 39%. Prevalence rates varying from 28% to 90% have been reported from southern India,14 Israel,15 Europe,2 Canada,16 Mexico17 and the USA.18 Each of these studies, however, reveals a unique set of CAM therapies selected from a wide spectrum, viz. diet modification, ayurveda, homoeopathy, chiropractic, acupuncture, copper bracelets and occult spiritual therapies. Geographical, social and cultural influences make the ascertainment of the absolute determinants of use of a specific CAM therapy difficult. The most common reason cited by our patients for using CAM was pain control, which is in contrast to the findings of Chandrashekara et al.14 whose patients used CAM because of the belief that modern medicine had no cure for RA and adverse reactions are rare with CAM.
   We observed that yoga asana and pranayama were more often used as ‘add ons’ to various disease-modifying agents while ayurveda and homoeopathy were used as alternatives. This could be due to the fact that these movement-based therapies are usually perceived by patients to be virtually free of any side-effects, thereby resulting in a higher proportion of patients continuing these therapies for longer periods. Our study was done in a referral centre and patients visiting us may not be representative of the patients in the community.
   Our study revealed a significant association of CAM use with age of the patient, which might be due to longer disease duration. Associations with either of these factors, however, were not detected by Rao et al.18 Other correlates of CAM use such as graduate education, female gender, higher economic status, which were reported by Rao et al.18 and Kaboli et al.,19 were not detected in our study. Interestingly, as was the case in the study by Rao et al.,18 our study too does not show any association between CAM use and activity of the disease.
   One important observation was that the majority of CAM users did not reveal this to the treating physician. The most common reason for this was the lack of enquiry by physicians. Physicians were unaware of as many as 72.7% of the CAM courses used by their patients. This has also been noted by other investigators.18,19 CAM use is an important consideration during initiation of disease-modifying drug therapies in patients, as several CAM therapies (especially oral drugs) may lead to adverse drug interactions. Currently, there is a paucity of information relating to the safety of CAM therapies.20 It is likely that in the absence of such knowledge, some of the adverse effects due to CAM are wrongly attributed to disease-modifying drugs by patients,
leading to inappropriate cessation of therapy. It is imperative for rheumatologists to routinely enquire about CAM use from their patients.

ACKNOWLEDGMENTS
We thank Drs S. N. Dwivedi and Vishnu Sreenivas, Department of Biostatistics, AIIMS for help with statistical analysis of the data. We also acknowledge and thank Wamiqur Rahman Gajdhar MB,BS student, for helping the team with data entry. TZ and SA were recipients of the Late Shri Amrut Mody Unichem BJD India Research Fellowship Award 2006.

REFERENCES