Painful Diabetic Neuropathy: Impact of an alternative approach

A total of 207 subjects with type 2 diabetes and PDN were recruited. Analgesics including anticonvulsants, tricyclic antidepressants, and nonsteroidal antiflammatory agents were permitted and equally distributed across groups, but no dosing adjustments were allowed. Subjects were stratified according to age, diabetes duration, diabetes control, and recruitment site. Written informed consent and ethical approval was obtained.

The study design was a randomized, semidouble–blind, placebo-controlled, 12-week trial. Subjects were randomized into one of three treatment groups (Reiki, mimic Reiki, or usual care) in a 1:1:1 fashion. Mimic practitioners were actors trained to mimic Reiki practitioners in style of practice. Reiki was applied as previously described (6). Practitioners were 75% female and of a similar age range to the patients. Patients randomized to Reiki or mimic Reiki underwent two treatments in the first week, followed by weekly treatments. Patients in the usual-care group were only assessed at the start and end of the 12-week period. Due to poor participant retention, this arm was discontinued after randomization of 26 subjects.

Therapy sessions of 25-min duration were conducted throughout the day with patients supine in a clinic room decorated to facilitate relaxation. The primary efficacy outcome was the McGill Pain Questionnaire (7). Secondary end points were the 6-min walk test (8), the Epidemiology of Diabetes Intervention and Complications quality-of-life questionnaire, the Well-Being Questionnaire, and the Diabetes Treatment Satisfaction Questionnaire (9–11).

Treatment group comparisons followed the intent-to-treat principle. Treatment groups were first compared using an overall (2 d.f.) test of difference. Pairwise comparisons were made using Tukey's test for multiple comparisons. Analyses were repeated after stratification in general linear regression models. Paired (pre/post) data were analyzed using Student's t test for the continuous variables. Participants without a final evaluation were dropped from the analysis. A significance level of 0.05 was used throughout the analyses.

Ninety-three subjects were randomized to Reiki group (mean ± SD age 66 ± 10 years, 63% male, A1C 8.7 ± 1.5%, and diabetes duration 9.2 ± 7.6 years), 88 to the mimic-Reiki group (age 65 ± 10 years, 61% male, A1C 8.5 ± 1.7%, and diabetes duration 9.8 ± 8.6 years), and 26 to the usual-care group (age 59 ± 10 years, 50% male, A1C 8.5 ± 1.5%, and diabetes duration 9.6 ± 6.7 years). Overall, 82, 75, and 69% of subjects completed the final evaluation in the Reiki, mimic-Reiki, and usual-care groups, respectively.

Compared with baseline, the total score for pain descriptors improved significantly (P < 0.05) for both the Reiki and mimic-Reiki treatments but not for the usual-care group (Table 1). However, final pain scores were not significantly different between groups. The trends toward improvement in the visual analogue scale and present pain intensity were not statistically significant. Treatment differences at 12 weeks were not significantly different either adjusted or unadjusted for baseline values and stratification factors.

In the Reiki, mimic-Reiki, and usual-care groups, 7, 3, and 1 pain descriptors improved, respectively (P < 0.05). At the final visit, walking distance improved by 12% (P = 0.005), 12% (P = 0.009), and 6% (P = not significant), respectively, in the Reiki, mimic-Reiki, and usual-care groups. There were no differences between treatments in the 12-week measures. Compared with baseline, at the final visit, no significant changes were detected in either the Well-Being Questionnaire or the Diabetes Treatment Satisfaction Questionnaire. The Epidemiology of Diabetes Intervention and Complications questionnaire detected borderline improvement within the Reiki group (P = 0.05).

Eight patients withdrew from the study due to serious adverse events, and one patient died. No event was related to study treatment.

Complimentary and alternative modalities are increasingly used in the treatment of diabetes or its complications (12), but often their efficacy remains unvalidated. This study therefore explored the efficacy of Reiki to control PDN as a complication of type 2 diabetes.

Global pain scores and walking distance improved from baseline in both the Reiki and mimic-Reiki groups. However, there were no significant differences between groups at the final visit. These results indicate that in this population, the application of Reiki was not effective for the treatment of PDN, and the responses observed in both Reiki and mimic-Reiki subjects are consistent with those expected for a clinical trial (13,14). However, limitations of this study that may have reduced the power to detect a difference between groups were the relatively low pain scores at baseline in all groups and the inability to retain usual-care subjects.

In conclusion, Reiki is no more effective than mimic-Reiki in decreasing pain perception and improving walking distance in subjects with PDN. However, the reduction of pain symptoms observed in both treatment groups is consistent with the concept that the formation of a “sustained partnership” between the health care provider and the patient can have direct therapeutic benefits (15).

This work was supported by National Institutes of Health Grants R01-F000727 and P50-AT-00011, the General Clinical Research Centers Grant M0–1RR-00042, the Juvenile Diabetes Research Foundation, Eli Lilly, and the American Diabetes Association.

We thank the Reiki practitioners, as well as Renu Mahajan and Robert Adwere-Boamah.