High-dose curcuminoids are efcacious in the reduction in symptoms and signs of oral lichen planus

Nita Chainani-Wu, DMD, MS, PhD,a,b Erin Madden, MPH,c Francina Lozada-Nur, DDS, MS, MPH,a and Sol Silverman, Jr, MA, DDSa San Francisco and Mountain View, California
Background: Curcuminoids are components of turmeric (Curcuma longa) that possess anti-inflammatory properties. Objective: We sought to study the efcacy of curcuminoids in controlling the signs and symptoms of oral lichen planus, at doses of 6000 mg/d (3 divided doses), and their safety at this dose. Methods: Twenty consecutive, eligible patients who consented were enrolled into this randomized, double-blind, placebo-controlled clinical trial in 2007 through 2008. Measurement of symptoms and signs of oral lichen planus using the Numerical Rating Scale (NRS) and the Modied Oral Mucositis Index (MOMI), respectively; complete blood counts; liver enzymes; C-reactive protein; and interleukin-6 levels was done at baseline and day 14. Two-sided P values are reported. Results: In the placebo group, the percentage changes from baseline in NRS (median [interquartile range] = 0.00 [29 to 16.7], P [ .99), erythema (0.00 [10 to 16.7], P = .98), ulceration (0.00 [0.00 to 26.7], P = .63), and total MOMI scores (3.2 [13 to 9.09], P = .95) were not statistically significant, whereas they were statistically significant in the curcuminoids group: NRS (22 [33 to 14], P = .0078); erythema (17 [29 to 8.3], P = .0078), ulceration (14 [60 to 0.00], P = .063), MOMI (24 [38 to 11], P = .0039). The curcuminoids group showed a greater reduction in clinical signs and symptoms as compared with the placebo group, measured by percentage change in erythema (P = .05) and total MOMI score (P = .03), and proportion showing improvement in NRS (0.8 vs 0.3, P = .02) and total MOMI score (0.9 vs 0.5, P = .05). Adverse effects were uncommon in both groups. Limitations: The small sample size resulted in limited power, particularly for multivariate analyses. Conclusions: Curcuminoids at doses of 6000 mg/d in 3 divided doses are well tolerated and may prove efcacious in controlling signs and symptoms of oral lichen planus. ( J Am Acad Dermatol 2012;66:752-60.) Key words: anti-inammatory; autoimmune; curcuminoids; herbal; oral lichen planus; turmeric.

ichen planus is an autoimmune, mucocutaneous condition. It commonly involves the oral mucosa, which may be the only affected site. Oral lichen planus (OLP) can cause oral discomfort, and is characterized by white reticular changes,

erythema, and ulcers.1 Curcuminoids are components of the root turmeric (Curcuma longa), and have antiinflammatory properties.2 Clinical studies of curcuminoids have evaluated its use in conditions such as rheumatoid arthritis, postsurgical inflammation, and

From the Department of Orofacial Sciences, School of Dentistry, University of California, San Franciscoa; private practice in oral medicine, Mountain Viewb; and Northern California Institute for Research and Education (NCIRE), Veterans Affairs Medical Center, San Francisco.c Supported by the Mount Zion Health Fund, through the Osher Center for integrative medicine, University of California, San Francisco. Sabinsa Corp (Dr Vladimir Badmaev, Piscataway, NJ) donated the curcuminoids and placebo capsules. Conflicts of interest: None declared. Accepted for publication April 28, 2011.

The findings from this study were presented at the Annual Meeting of the American Academy of Oral Medicine, Santa Ana Pueblo, NM, April 13-17, 2010. Reprint requests: Nita Chainani-Wu, DMD, MS, PhD, University of California, San Francisco, 521 Parnassus Ave, C646, Box 0658, San Francisco, CA 94143-0658. E-mail: nita.wu@ucsf.edu. Published online September 12, 2011. 0190-9622/$36.00 2011 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2011.04.022

752

J AM ACAD DERMATOL
VOLUME 66, NUMBER 5

Chainani-Wu et al 753

chronic uveitis.2 Therefore, in 2003 through 2004 we obtained from the Food and Drug Administration. conducted a randomized, double-blind, placeboThis study was registered on clinicaltrials.gov in controlled trial of 2000 mg/d of curcuminoids in 33 September 2007, identifier: NCT00525421. patients with OLP.3 We ended this study at the first Eligible patients were those older than 21 years; interim analysis for futility. Data from this earlier study with a current clinical presentation of atrophic or and observations from open-label use of curcumierosive OLP, symptom score for OLP between 3 and noids in patients with OLP indicated that a starting 8 at enrollment (range of scale, using the Numerical dose of 2000 mg/d of curcuRating Scale [NRS]: 0 [no oral minoids was insufficient. discomfort] to 10 [worst CAPSULE SUMMARY Therefore we conducted the imaginable oral discomfort]); current randomized conand who had not used topiCurcuminoids are components of trolled trial using a higher cal or systemic glucocorticosturmeric (Curcuma longa) that possess dose of curcuminoids (6000 teroids for at least 2 weeks. anti-inflammatory properties. mg/d in 3 divided doses) in 20 Exclusion criteria included In a randomized, double-blind, placebopatients with OLP. pregnancy, lactation, paThe primary objectives of controlled 2-week study of curcuminoids tients on long-term glucocorthis study were to evaluate in 20 patients with oral lichen planus, a ticosteroid therapy, current the efcacy of curcuminoids greater reduction in symptoms and signs orthodontic treatment, and at a dose of 6000 mg/d in 3 was observed in the curcuminoids group history of gastroesophageal divided doses in controlling as compared with the placebo group. reux disease, gastric ulcers, the signs and symptoms of duodenal ulcers, gallstones Adverse effects included diarrhea but OLP, and the safety and tol(curcumin may cause gastric were uncommon. erability of curcuminoids at irritation and curcumin can Curcuminoids at doses of 6000 mg/d in 3 this dose (specic aim 1). stimulate gall bladder condivided doses are well tolerated and may In addition, we planned to tractions4), or elevated liver prove efficacious in controlling signs and evaluate the association of enzymes above 2.5 times the symptoms of oral lichen planus. serum C-reactive protein upper limit of normal (cur(CRP) and serum interleukin cumin can cause hepatotox(IL)-6 with clinical severity of OLP at baseline and icity in some species such as mice and rat2). over time (specic aim 2), and the effect of curcuThe primary outcome of the study was change in minoids on these proteins (specic aim 3). We also symptom scores from baseline to 2-week follow-up. planned to evaluate whether subgroups of patients Secondary end points included changes in clinical with varying responses to curcuminoids could be signs, CRP, and IL-6 from baseline, and occurrence of identied by their baseline serum CRP or serum IL-6 adverse effects. levels (specic aim 4), and whether changes in these Subjects were recruited from the oral medicine proteins could explain the mechanism of action of clinic at UCSF. Patients were screened by a review of curcuminoids in OLP (specic aim 5). the medical history, medications used, current symptom score (for OLP), and an oral examination. At this METHODS time potential subjects were identied, written inStudy design formed consent was obtained and consenting paThis was a phase II randomized double-blind, tients were scheduled for a blood draw on the same placebo-controlled clinical trial. day. Those with liver enzymes within 2.5 times the upper limit of normal were enrolled and randomized. The study medication was sent to them by Study population overnight mail, allowing a 12-day course of study A total of 23 patients with symptomatic OLP medication taken before their 2-week follow-up visit. presenting to the oral medicine clinic at the A consecutive sample of eligible patients was University of California, San Francisco (UCSF) beenrolled into the study. Subject enrollment and tween October 2007 and November 2008 were follow-up is outlined in Fig 1. screened for this study. Twenty of these patients met the eligibility criteria, signed a written consent form to participate, and were enrolled and randomIntervention ized. Three patients had elevated liver enzymes at The study medication was a standardized extract baseline and were excluded (Fig 1). This study was of turmeric, which contains at least 95% curcumiapproved by the UCSF Institutional Review Board. noids, called curcumin C3 complex. The turmeric An investigational New Drug Approval number was rhizomes used for the manufacture of curcumin C3
d d d d

754 Chainani-Wu et al

J AM ACAD DERMATOL

MAY 2012

Abbreviations used: CRP: IL: MOMI: NRS: OLP: UCSF: C-reactive protein interleukin Modied Oral Mucositis Index Numerical Rating Scale oral lichen planus University of California, San Francisco

complex were cultivated in Salem District, Tamil Nadu, India. Sabinsa Corp (Piscataway, NJ) holds a patent (US Patent 5,861,415; January 19, 1999) on the Bioprotectant composition, method of use and extraction process of curcuminoids for curcumin C3 complex. According to the information provided by Sabinsa Corp, the purity of curcuminoids in curcumin C3 complex is as follows: curcumin between 70% and 80%, demethoxycurcumin between 15% and 25%, and bisdemethoxycurcumin between 2.5% and 6.5%, depending upon the batch. Study medication tablets with 95% curcuminoids and identical, placebo tablets (microcrystalline cellulose, dicalcium phosphate, PVPK 30, sodium starch glycolate, magnesium stearate, OpaDry orange coating, Colorcon, Harleysville, PA) were provided by Sabinsa Corp. A certicate of analysis was provided by Sabinsa Corp for batch No. CF71426 (used in the study) that showed potency assayed by high-performance liquid chromatography and contaminant levels (heavy metals, arsenic, and lead) to be in the acceptable range. This was conrmed by an independent laboratory analysis obtained from American Analytical Chemistry Laboratories, Champaign, IL. This independent laboratory analysis also included testing for microbes (yeast and molds, Escherichia coli, Staphylococcus aureus, Pseudomonas aeruginosa, and a Salmonella screen), which were all negative. Randomization and blinding A stratied, blocked randomization (block size of 6) was used. Stratication was based on symptom score at baseline, with two strata (symptom score ranging from 3-5 and from [5-8). The UCSF pharmacy packaged the curcuminoids and identical placebo tablets in plastic medication containers and generated the randomization sequence using the random number generator in Microsoft Excel (Microsoft Corp, Seattle, WA). Both participants and investigators were blinded to the treatment assignment. Measurement of response variables (1) NRS. The range is from 0 to 10 (0 = no symptoms and 10 = worst imaginable symptoms).

The NRS has been validated for measurement of symptoms associated with OLP.5 At the baseline visit and follow-up visit, symptom intensity over the previous week was recorded. On the follow-up visit, the study participant was shown the symptom score from the baseline visit, before they completed the NRS. (2) Modied Oral Mucositis Index (MOMI). Clinical signs of OLP were measured at the baseline and 2-week follow-up visit, using this semiquantitative scale that we previously developed and validated for measurement of clinical signs of OLP.5 An oral examination was conducted, and the atrophic and erosive changes were quantified based on severity and the number of sites involved. An intensity score for erythema ranging from 0 to 3 was used (0 = normal, 1 = mild erythema, 2 = moderate erythema, 3 = severe erythema). The score for ulcerations was based on area of ulceration (0 = no ulcerations, 1 = between 0-0.25 cm2, 2 = between 0.25-1 cm2, 3 = $ 1 cm2). Sixteen different oral sites were evaluated including, buccal mucosa (right and left), labial mucosa (upper and lower), lateral aspect of tongue (right and left), dorsum of tongue (right and left), ventral tongue and floor of mouth (right and left), maxillary gingiva (right and left), mandibular gingiva (right and left), soft palate, and hard palate. The scores for erythema and ulcerations was obtained by summing the respective scores for these 16 sites and the total score for clinical signs was obtained by summing the erythema and ulceration scores. (3) Side effects. Side effects were recorded using a 10-item questionnaire at the 2-week followup, and a complete blood cell count and liver enzymes, aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatase measurement at baseline and 2-week follow-up. (4) CRP and IL-6. CRP and IL-6 were measured at baseline and 2-week follow-up. All blood draws and blood tests were done at Quest Diagnostics Laboratory, San Francisco, CA. (5) Bleeding index. This index was measured by running a probe along the facial gingival sulcus of all teeth in a quadrant, following which each tooth was graded on a scale of 0 to 4 for presence of bleeding. This was repeated for each quadrant. The sum of the scores for all teeth was divided by the total number of teeth to calculate the bleeding index.6 Statistical analysis 1. Descriptive statistics including means, SD, medians, minimum, maximum, and percentages were used to summarize all variables (Table I).

J AM ACAD DERMATOL
VOLUME 66, NUMBER 5

Chainani-Wu et al 755

Fig 1. Subject enrollment and follow-up.

2. Changes in primary and secondary outcomes within and between treatment arms (specific aim 1 and specific aim 3). The primary outcome, percentage change in symptoms scores, and secondary outcomes of percentage change in clinical signs, CRP, and IL-6 were compared across treatment arms via the Mann-Whitney U test. For analyses of percent change from baseline, variables with any zero values at baseline had a constant (0.5) added to all observations to avoid infinite percent changes; this included the following variables: ulceration and bleeding index. The significance of the change from baseline within each treatment arm was evaluated using the Wilcoxon signed rank test. Pearson x 2 tests were done to test for the difference in proportion of participants in the curcuminoids groups versus the placebo group who had an improvement

(reduction) in NRS, MOMI, erythema, and ulceration scores. 3. Occurrence of side effects in the two groups was compared (specific aim 1) using the Pearson x2 test and the Mann-Whitney test. 4. Association of serum CRP and serum IL-6 with clinical severity of OLP at baseline and over time (specific aim 2). Spearman rho was calculated to evaluate correlations between baseline CRP and IL-6 with baseline symptoms, signs, and bleeding index scores. Multiple linear regression analyses of these associations were conducted controlling for baseline bleeding index. Spearman rho was also calculated to evaluate please check formatting of this paragraph correlations between changes in CRP and IL-6 with changes in symptoms, signs, and bleeding index scores; and

756 Chainani-Wu et al

J AM ACAD DERMATOL

MAY 2012

Table I. Baseline characteristics of study participants

Variable Placebo, n = 10 Curcuminoids, n = 10

plus MOMI. The percentage change in CRP and IL-6 from baseline was compared in responders and nonresponders. 7. Blinding. The kappa score for agreement between subject guess and correct assignment beyond chance and between the investigators guess and correct assignment beyond chance were calculated. The x2 tests for proportion correct for subject and investigator guesses were calculated. 8. Compliance. Percent compliance was calculated as: [(number of pill taken/number of pills to be taken) * 100]. 9. Sample size. The sample size of 20 was based on expected and actual enrollment of study subjects over a 12-month time period. Because of feasibility reasons the enrollment was stopped at 20 subjects. 10. All reported P values are two-sided. Analyses were performed using SAS Version 9.1 (SAS Institute, Cary NC).

Age (mean, SD) 56.2, 11.7 Male 5 (50.0%) Female 5 (50.0%) Duration of oral lichen 83.9, 105.3 planus (mean, sd) Duration of symptomatic 44.9, 61.0 oral lichen planus (mean, sd) Symptoms at baseline (Numerical Rating Scale): (1) 3-5 6 (60.0%) (2) 5.5-8 4 (40.0%) Ulcerations at baseline: (1) Yes 4 (40.0%) (2) No 6 (60.0%) Number of subjects who used medications used in the past for oral lichen planus: (1) Topical Steroids 7 (70.0%) (2) Prednisone 5 (50.0%) (3) Azathioprine 1 (10.0%) (4) Any medication used 8 (80.0%) (including above)

60.8, 8.6 2 (20.0%) 8 (80.0%) 69.5, 128.1 62.3, 110.5

5 (50.0%) 5 (50.0%) 3 (30.0%) 7 (70.0%)

8 6 1 9

(80.0%) (60.0%) (10.0%) (90.0%)

RESULTS
Please note that the results presented below are ordered to correspond to the numbered subsections in Statistical analysis. 1a. Baseline characteristics Baseline characteristics of the curcuminoids group and placebo group were similar (Table I). 1b. Levels of CRP and IL-6 at baseline and 2 weeks The range of values for CRP at baseline was 0.1 to 2.1 mg/dL, with a median value of 0.2 mg/dL. At 2 weeks, these values were: range 0.1 to 2.5 mg/dL, with a median value of 0.3 mg/dL. Seventeen of 19 (85%) patients at baseline and 15 of 19 (65%) at 2 weeks had values within the normal range of 0 to 0.8 mg/dL, whereas the remaining had elevated values. The range of values for IL-6 at baseline was 0.5 to 6.8 pg/mL, with median value of 1.5 pg/mL. At 2 weeks, these values were: range 0.2 to 16.5 pg/mL, with median value of 1.6 pg/mL. Nineteen of 20 (95%) patients at baseline and 16 of 20 (70%) at 2 weeks had values within the laboratory normal range of 0.31 to 5.00 pg/mL; one patient at baseline and 3 patients at 2 weeks had elevated values; whereas one patient at 2 weeks had a value below the lower limit of normal range. 2a. Effect of curcuminoids on symptoms and signs of OLP Percentage changes from baseline in symptoms and signs as measured by NRS and MOMI are

multiple regression analyses were done, controlling for baseline bleeding index. In addition, Spearman correlations for baseline CRP with baseline IL-6 and changes in CRP with changes in IL-6 were calculated. 5. Baseline CRP and baseline IL-6 as predictors of treatment responses (specific aim 4). The distribution of CRP and IL-6 values was nonnormal, and they were log transformed (log base 2) for these analyses; the coefficients are therefore the estimated effect per 2-fold increase in CRP or IL-6. To evaluate effect modification by baseline CRP and IL-6 levels on the association of treatment group and changes in symptoms and signs, linear regression models that included the independent variables treatment group, log2CRP and their interaction term were run with each of the following outcome variables: (1) change in NRS, (2) change in erythema, (3) change in ulceration, and (4) change in MOMI. The same models were built using log2IL-6 in place of log2CRP. 6. The role of CRP and IL-6 in the mechanism of action of curcuminoids (specific aim 5). Participants assigned to the curcuminoids group were classified as responders and nonresponders, with responder defined as being at or below the median percentage change in NRS

J AM ACAD DERMATOL
VOLUME 66, NUMBER 5

Chainani-Wu et al 757

Table II. Baseline and follow-up values, and percentage changes from baseline in symptoms and signs of oral lichen planus and C-reactive protein and interleukin-6 values

CRP, C-reactive protein; IL, interleukin; IQR, interquartile range; MOMI, Modified Oral Mucositis Index; NRS, Numerical Rating Scale. *Based on Wilcoxon signed rank test. y Based on Mann-Whitney U test. z Ulceration: constant of 0.5 was added to all baseline and 2-wk values because of existence of some zero values for this variable.

Curcuminoids Baseline Median (IQR)

5.25 7.75 2 9.5 0.25 2.175

presented in Table II and Fig 2. The percentage changes from baseline in NRS (median [interquartile range] = 0.00 [29 to 16.7], P [ .99), erythema (0.00 [10 to 16.7], P = .98), ulceration (0.00 [0.00-26.7], P = .63), and total MOMI scores (3.2 [13 to 9.09], P = .95) were not significant in the placebo group, whereas they were statistically significant in the curcuminoids group in NRS (22 [33 to 14], P = .0078), erythema (17 [29 to 8.3], P = .0078), and MOMI scores (24 [38 to 11], P = .0039). The curcuminoids groups showed a greater reduction in symptoms measured by percentage change in NRS as compared with the placebo group (P = .09) and a greater reduction in signs as measured by percentage change in erythema (P = .05), ulceration (P = .08), and total MOMI score (P = .03) as compared with the placebo group; in addition, at the follow-up visit, a greater proportion of participants in the curcuminoids group showed improvement in symptoms and signs of OLP as compared with the placebo group as measured by NRS (0.8 vs 0.3, P = .024), MOMI (0.9 vs 0.5, P = .05), erythema (0.8 vs 0.4, P = .06), and ulceration (0.5 vs 0.2, P = .15), although not all these P values reached statistical significance. 2b. Effect of curcuminoids on CRP and IL-6 Percentage changes from baseline in CRP and IL-6 levels were not statistically signicant in the placebo and curcuminoids group, nor was the difference in percentage change in CRP and IL-6 between the curcuminoids and placebo group (Table II). One participant had a high outlier value for percentage change in IL-6 (428%), and this participant had dramatic improvements in symptoms (50% percentage change in NRS) and signs (44.4% change in ulceration score) of OLP at the follow-up visit. 3. Adverse effects of curcuminoids Curcuminoids were well tolerated. Occurrence of adverse effects as reported by study participants were similar in the curcuminoids and placebo groups. Complete blood count values and liver enzymes evaluated as proportion showing change from normal to abnormal (Table III) and as continuous values (percentage change from baseline) showed no statistically significant differences between placebo and curcuminoids groups. 4a. Association of OLP severity with CRP and IL-6 levels at baseline Table IV shows Spearman correlation coefficients for baseline CRP and IL-6 with baseline bleeding index and symptoms and signs of OLP.

P valuey Rank P value* Curcuminoids Percent change from baseline Median (IQR) Curcuminoids 2-wk Median (IQR) Rank P value* Placebo Percent change from baseline Median (IQR) Placebo 2-wk Median (IQR) Placebo Baseline Median (IQR) Variable

NRS Erythema Ulcerationz MOMI CRP IL-6

5 7 1.5 9.5 0.2 1.32

(4.5-7) (5-11.5) (0.5-5.5) (6-15.5) (0.1-0.3) (0.99-1.73)

5 7.25 1.5 9 0.15 1.42

(4.5-6) (4.5-9) (1-4.5) (5.5-12) (0.1-0.4) (0.93-1.59)

0.00 0.00 0.00 3.2 33.3 1.6

(29 to 16.7) (10 to 16.7) (0.00-26.7) (13 to 9.09) (0.00-50.0) (6.1 to 3.17)

1.00 .98 .63 .95 .13 .57

(3.5-7) (6-9.5) (0.5-6.5) (7-17) (0.1-0.6) (0.99-4)

3.75 6.5 1.25 7.75 0.4 1.65

(3-5) (4-8.5) (0.5-4) (5.5-11) (0.2-1.1) (1.32-5.72)

22 17 14 24 0.00 20.3

(33 (29 (60 (38 (21 (16

to to to to to to

14) 8.3) 0.00) 11) 19.0) 60.6)

.0078 .0078 .063 .0039 1.00 .32

.099 .052 .082 .031 .19 .089

758 Chainani-Wu et al

J AM ACAD DERMATOL

MAY 2012

Table III. Adverse events

Adverse events Placebo Curcuminoids P value

Diarrhea Constipation Abdominal pain Heartburn Nausea Fatigue Insomnia Mood changes Bloating Any reported side effect including above Liver enzymes (AST, ALT and alkaline phosphatase): change from normal to abnormal value

1 1 0 0 0 0 0 0 0 2

(10.0%) (10.0%) (0.0%) (0.0%) (0.0%) (0.0%) (0.0%) (0.0%) (0.0%) (20%) 0

2 1 1 1 1 0 0 0 0 4

(20.0%) (10.0%) (10.0%) (10.0%) (10.0%) (0.0%) (0.0%) (0.0%) (0.0%) (40%) 2*

.53 1.00 .30 .30 .30 e e e e .33 .16

ALT, Alanine aminotransferase; AST, aspartate aminotransferase. *ALT changed from 40 to 44, and from 59 to 64.

Baseline CRP levels were inversely correlated with baseline NRS scores (r = 0.445, P = .05, N = 20), and this inverse association remained significant after controlling for baseline bleeding index (B = 0.16 [95% confidence interval 0.29 to 0.03], P = .01) in multiple regression analyses. Associations with baseline erythema, ulceration, and total MOMI were minimal and nonsignificant, both in bivariate analysis and after controlling for bleeding index scores. Baseline IL-6 levels were positively correlated with baseline erythema scores (r = 0.449, P = .04, N = 20), whereas the coefficient in the regression model controlling for bleeding index was B = 0.25 (95% confidence interval 0.05 to 0.55), P = .09. Associations with baseline NRS, ulceration, and total MOMI were minimal and nonsignificant, both in bivariate analysis and after controlling for bleeding index scores. 4b. Association of changes in OLP severity with changes in CRP and IL-6 Bivariate Spearman correlations that show inverse associations of change in CRP and change in IL-6 with change in symptoms and signs and bleeding

=
Fig 2. Boxplots showing percentage change in symptoms and signs of oral lichen planus (A and B) and C-reactive protein (CRP) and interleukin (IL)-6 levels (C) in curcuminoids and placebo group. Bottom and top of box are

lower and upper quartiles, and band near middle of box is median. Ends of whiskers are lowest datum still within 1.5 interquartile range (IQR) of lower quartile, and highest datum still within 1.5 IQR of upper quartile. Outliers are represented by small circle. MOMI, Modified Oral Mucositis Index; NRS, Numerical Rating Scale.

J AM ACAD DERMATOL
VOLUME 66, NUMBER 5

Chainani-Wu et al 759

Table IV. Association of baseline and change scores for C-reactive protein and interleukin-6 with bleeding index, symptoms and signs of oral lichen planus, N = range 18 to 20 because of missing values (Spearman rho)
Baseline CRP Baseline IL-6 Change in CRP Change in IL-6

Baseline Baseline Baseline Baseline Baseline Baseline Baseline Change Change Change Change Change Change Change

bleeding index NRS erythema ulceration MOMI CRP IL-6 in in in in in in in bleeding index NRS erythema ulceration MOMI CRP IL-6

0.225 0.445 0.337 0.056 0.190

(P (P (P (P (P 1

= = = = =

.36) .056) .16) .82) .44)

0.247 .109 0.449 0.207 0.362 0.352

(P = (P = (P = (P = (P = (P = 1

.29) .65) .047) .38) .12) .14) 0.184 0.244 0.040 0.039 0.083 (P (P (P (P (P 1 = = = = = .46) .33) .87) .88) .74) 0.080 0.359 0.259 0.256 0.405 0.107 (P (P (P (P (P (P 1 = = = = = = .74) .12) .27) .28) .076) .67)

CRP, C-reactive protein; IL, interleukin; MOMI, Modified Oral Mucositis Index; NRS, Numerical Rating Scale.

index scores are presented in Table IV. However, these associations were not statistically significant. Linear regression analysis controlling for bleeding index were also run, which did not show any significant associations. 5. Effect of baseline CRP and baseline IL-6 on treatment responses The interaction terms in the models built to evaluate effect of baseline CRP and baseline IL-6 on treatment responses (signs and symptoms) were not signicant in any of these models. These models were rerun with change in bleeding index, a possible confounder, included in each model with the same results. 6. Treatment responses and changes in CRP and IL-6 in the curcuminoids group Responders had median (range) percentage change in CRP = 26.2 (21.4 to 500, N = 5), whereas nonresponders had median (range) percentage change in CRP = 0.0 (50 to 0, N = 5), with P value for the difference in change = 0.11. Responders had median (range) percentage change in IL-6 = 26.0 (16.4 to 428, N = 5), whereas nonresponders had median (range) percentage change in IL-6 = 13.0 (68.5 to 175.0, N = 5), with P value for the difference in change = 0.40. 7. Assessment of blinding At the last follow-up visit, the investigator (N. C-W.) and 19 of the 20 study participants indicated whether they believed participants had been assigned to the treatment group or the placebo group. In all, 50% of subjects correctly guessed their treatment assignment and the investigator correctly

guessed the treatment assignment of 70% of the study subjects. The kappa score for agreement between subject guess and correct assignment beyond chance was 0.06 (0.3, 0.4), and the kappa score for agreement between the investigators guess and correct assignment beyond chance was 0.4 (0.006, 0.7). The x 2 tests for proportion correct for subject guesses (P [ .999) and investigator guesses (P = .07) as compared with the proportion 0.5 were not significant. These results suggest adequate blinding of the study medication. 8. Compliance Compliance was measured by pill counts. Nine of 10 patients in the placebo group had more than 86% compliance (range 70.8%-100%); 9 of 10 patients in the curcuminoids group had over 91% compliance (range 85.4%-100%).

DISCUSSION
This is the rst study demonstrating the efcacy of curcuminoids in the reduction of symptoms and signs of OLP. This study used a dose of curcuminoids of 6000 mg/d in 3 divided doses for 12 days whereas our previous randomized controlled trial that was ended early for futility used a dose of 2000 mg/d in two divided doses for 7 weeks.3 The results therefore suggest that the higher dose of 6000 mg/d should be used when initiating treatment for OLP. Curcuminoids at this dose were well tolerated, with diarrhea being the most frequently reported side effect. OLP responds to systemic and topical glucocorticosteroids, which are the main currently prescribed treatments for OLP.1 It is therefore reasonable to expect that curcuminoids may help

760 Chainani-Wu et al

J AM ACAD DERMATOL

MAY 2012

in other autoimmune diseases and conditions where glucocorticosteroids have applications. Because of the better safety profile of curcuminoids, there are significant advantages to the use of curcuminoids versus the use of systemic glucocorticosteroids. Therefore, these results suggest that clinical trials evaluating the use of curcuminoids in other autoimmune conditions should be considered. The serum markers CRP and IL-6 were measured to describe their levels in patients with OLP, to study the effect of curcuminoids on these markers, to identify subgroups of patients with OLP who may show varying responses to curcuminoids, and to understand the role of CRP and IL-6 in the pathophysiology of OLP and in the mechanism of action of curcuminoids. The main limitation of the analyses pertaining to CRP and IL-6 was the lack of power because of the small sample size. The results from these analyses, although not statistically signicant, provide some information on direction of responses, which may help guide the design of future studies. Most participants at baseline had low serum levels of CRP and IL-6. The difference in the changes in these markers after administration of curcuminoids as compared with placebo was not statistically signicant. Studies on IL-6 have found evidence of antiinammatory activity7-10 and proinflammatory activity11,12 and a complete picture of its role in inflammation has not yet been elucidated. The significance of the correlations of CRP and IL-6 with signs and symptoms of OLP at baseline in Table IV are unclear. Table IV also shows the correlation of changes in CRP and IL-6 with changes in symptoms and signs of OLP. Most of these are inverse correlations, suggesting an increase in these markers with reduction of symptoms and signs of OLP at 2 weeks; also the percentage change in IL-6 among the responders was greater than that seen in the nonresponders within the curcuminoids group. However, the P values for these tests did not reach statistical significance. The findings from these analyses suggest a possible role for IL-6 in the mechanism of action of curcuminoids and in the pathophysiology of OLP, and suggest the need to evaluate this in future studies. In conclusion, this study suggests the efcacy of high-dose curcuminoids in reduction of symptoms and signs of OLP. The ndings are consistent with

and add to the safety data in the literature showing that oral administration of curcuminoids is safe and well tolerated.
We would like to thank all the study participants, Kornelia Collins for data entry and clerical assistance, Mohamed Shameem for creating the Microsoft Access database, Dr Peter Loomer and Dr Gary Armitage for discussions relating to the gingival bleeding measurements, and Dr Piri Veluppillai and Dr Chaitra Bhat for referring patients to the study.

REFERENCES 1. Chainani-Wu N, Silverman S Jr, Lozada-Nur F, Mayer P, Watson J. Oral lichen planus: patient profile, disease progression and treatment responses. J Am Dent Assoc 2001; 132:901-9. 2. Chainani-Wu N. Safety and anti-inflammatory activity of curcumin: a component of turmeric (Curcuma longa). J Altern Complement Med 2003;9:161-8. 3. Chainani-Wu N, Silverman S Jr, Reingold A, Bostrom A, McCulloch C, Lozada-Nur F, et al. A randomized, doubleblind, placebo-controlled clinical trial of curcuminoids in oral lichen planus. Phytomedicine 2007;14:437-46. 4. Rasyid A, Lelo A. The effect of curcumin and placebo on human gall-bladder function: an ultrasound study. Aliment Pharmacol Ther 1999;13:245-9. 5. Chainani-Wu N, Silverman S Jr, Reingold A, Bostrom A, LozadaNur F, Weintraub J. Validation Of visual analogue scale, numeric rating scale, change in symptoms scale and modified oral mucositis index for measurement of symptoms and signs of oral lichen planus. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2008;105:51-8. 6. Newbrun E. Indices to measure gingival bleeding. J Periodontol 1996;67:555-61. 7. Tilg H, Trehu E, Atkins MB, Dinarello CA, Mier JW. Interleukin-6 (IL-6) as an anti-inflammatory cytokine: induction of circulating IL-1 receptor antagonist and soluble tumor necrosis factor receptor p55. Blood 1994;83:113-8. 8. Xing Z, Gauldie J, Cox G, Baumann H, Jordana M, Lei XF, et al. IL-6 is an antiinflammatory cytokine required for controlling local or systemic acute inflammatory responses. J Clin Invest 1998;101:311-20. 9. Steenseberg A, Fisher CP, Keller C, Meller K, Pedersen BK. IL-6 enhances plasma IL-1ra, IL-10, and cortisol in humans. Am J Physiol Endocrinol Metab 2003;285:E433-7. 10. Petersen AMW, Pedersen BK. The role of IL-6 in mediating the anti-inflammatory effects of exercise. J Physiol Pharmacol 2006;57:43-51. 11. Kamimura D, Ishihara K, Hirano T. IL-6 signal transduction and its physiological roles: the signal orchestration model. Rev Physiol Biochem Pharmacol 2003;149:1-38. 12. Gabay C. Interleukin-6 and chronic inflammation. Arthritis Res Ther 2006;8(Suppl):S3.