JRHS 2009; 9(2): 19-24

Copyright © Journal of Research in Health Sciences

Assessment of Marhame-Mafasel Pomade Effect on Knee Osteoarthritis with Non-Compliance

Soltanian AR (MSc)a, Faghihzadeh S (PhD)a, Mehdibarzi D (MD)b, Gerami A (PhD)c, Nasery M (MD)d, Cheng J (PhD)e

a Department of Biostatistics, School of Medical Sciences, Tarbiat Modares University, Tehran, Iran

b Department of Orthopedics, Faculty of Medical Sciences, Shahed University, Tehran, Iran

c Department of Mathematical Statistics, School of Mathematics, Statistics and Computer Sciences, Tehran Uni­versity, Iran

d Department of Pharmacology, Faculty of Medical Sciences, Shahed University, Tehran, Iran

e Department of Biostatistics &Epidemiology, University of Florida, Gainesville, U.S.A

*Corresponding author: Dr Soghrat Faghihzadeh, E-mail: faghihzadehs@yahoo.com.

Received: 2 Septamber 2009; Accepted: 23 November 2009


Background: Osteoarthritis is the most prevalent chronic non-infective joint arthritis. In the present study, the effect of new herbal pomade (Marhame-Mafasel) on knee osteoarthritis was investigated in a randomized trial. The objective of this study was to assess efficacy of Marhame- Mafasel pomade, which was consisted of several medic herbs like Arnebia euchroma and Martricaria chamomilla in pri¬mary osteoarthritis of the knee with non-compliance.

Methods: The 22 crossover trial enrolled 42 osteoarthritis patients (Marhame-Mafasel versus pla¬cebo) in 2006. The instrument of data collection was Western Ontario and McMaster Universities (WOMAC) LK3.1 standard questionnaires. We used conditional estimation to adjust non-compliance ef¬fect.

Results: The participants in each group were 21 patients. About 30 (71.4%) were female. The partici¬pants were between 40-76 years old. Positive analgesic effect of herbal pomade “Marhame-Mafasel” on knee osteoarthritis severity was considerable (P< 0.01). After adjusting results to compliance level, the estimators were sharper than crude results.

Conclusion: Herbal joint pomade "Marhame-Mafasel" has significant positive analgesic effect on pri¬mary knee osteoarthritis.

Keywords: Osteoarthritis, Herbal medicine, Compliance, Randomized trials


Osteoarthritis was known as degenerative joint disease occurs when the cushiony carti­lage be­tween two bones becomes worn down, and the bones begin to rub against each other in the knee joint (the area where two bones come to­gether) (1). Osteoarthritis of the knee often leads to pain, swelling, limi­tation in range of motion, stiffness, or the formation of bone spurs (tiny growths of new bone) (1). Osteo­arthri­tis is the most prevalent chronic non-in­fec­tive joint ar­thri­tis. Approximately 25% of people at age 55 yr or above have daily knee pain in (2). There is a significant positive cor­relation be­tween age and osteoarthritis of the knee (3). The preva­lence of this disease in women is greater than men (4). Prevalence of osteo­arthri­tis of the knee in Ame­r­ica is ap­proximately 0.9% (1.2% in women and 0.4% in men) (5). Osteoarthritis of the knee is one of the main leading causes of impaired mo­bil­ity in the elderly people (6). Many pa­tients with knee pain have limitations in their physi­cal func­tions, which prevented them from en­gag­ing in their usual daily activities.

Drugs more frequently used in osteoarthritis are analgesics, supporter of cartilage, steroid, and non-steroidal anti-inflammation drugs (NSAD). In addition, there are many pharma­cological, supportive, and surgical in­terventions, which de­pend on the disease se­verity. The disease is chronic; hence, drugs used locally are pre­ferred due to less complica­tion. As steroid and non-steroidal anti-inflammation drugs have sys­temic side ef­fects like digestive and renal impair­ment, they should be used carefully (7, 8). Local drugs like pomade, cream, gel, etc. are sim­ply used. Thus, preparing pomade to reduce pain and dis­ability of patients is very impor­tant. Despite the long history of herbal medi­cine in Iran, a few studies were carried out to investigate the ef­fect of herbal medication on osteoarthritis.

In order to reduce pain in patients suffering from osteoarthritis, the effect of new herbal po­made on osteoarthritis of the knee was in­ves­ti­gated in a double-blinded crossover trial.


Double-blinded crossover randomized trial on efficacy of herbal joint pomade (EHJP)

The EHJP study(9) conducted a double blinded placebo controlled randomized cross­over trial in­volving 42 osteoarthritis pa­tients aged 40 to 80 yr who had explicit symptoms of arthritis dis­ease to investigate the effect of herbal joint pomade “Marhame-Mafasel” (EHJP) on knee os-­teoarthritis, which participants drown from pa­tients attend­ing the Clinic of Mostafa- Khomeini Hospital in 2006. Pomade "Marhame-Mafasel" (MM) consisted of several medic herbs (like Arnebia euchroma and Martri­caria chamomilla) and was made by Pharma­cology Division of Shahed University, Iran. MM pomade and pla­cebo were inserted in the similar tubs. Pa­tients with acute knee ar­thritis or secondary os­teo­arthritis were ex­cluded from the study. The pro­to­col was ap­proved by the Research Ethics Committee at Shahed University, Tehran. Be­fore starting of the study, participants signed the in­formed consent forms according to Hel­sinki Declaration rule. Then a computer ran­dom number generator was used to allocate par­tici­pant to either placebo or treatment groups. Pa­tients used locally either MM po­made or pla­cebo 3 times a day for 3 wk. After 3 wk, sub­jects were as­sessed using checklist. Subjects were evalu­ated based on three characteristics including: a) pain score ranged from 0 (no pain) to 100 (extreme pain); b) physical function score ranged from 0 (no difficulty) to 100 (ex­treme dif­ficulty); and c) stiffness score ranged 0 (no stiffness) to 100 (extreme stiff­ness) at the end of both periods. These charac­teristics were meas­ured using Western Ontario and McMaster uni­ver­sities (WOMAC) checklist. After 1 wk wash out period, participants received alter­na­tive in­ter­vention in period II. In this study, we had two sequences: AB (MM pomade fol­lowed by pla­cebo) and BA (placebo fol­lowed by MM). The participants were known as com­pli­ance if they had consumed 50% or more of the as­signed po­made, other­wise they were known as non-com­pli­ance. In this study, the non-com­pliance and com­pli­ance distribution was showed in Ta­ble 1.

Table 1: The Distribution of Compliance status (two levels) in two periods of two sequences

Period I

Period II



Compliance Level


Proportion of



Proportion of






























* In AB sequence MM pomade, A, followed by placebo, B; and in BA sequence placebo followed by MM pomade.

Statistical model and analysis

We used principle component analysis and con­sidered a new outcome that was a linear com­pound of three characteristics including pain, physical function, and stiffness scores. The new outcome was named osteoarthritis in­tensity score ranged 0(no intension) to 100(extreme inten­sity). In the EHJP study, the observed com­pliance status of each sub­ject was classified to a binary variable (1 or 0) based on the amount of pomade in tubes taken by the subject. One subject was consid­ered to comply with the as­signed treat­ment (compliance= 1) if more than 50% of the pomade in the tubes was taken. Other­wise, the subject's observed compli­ance to the assigned treatment was classified as 0. Un­der complete data assumption (9), poten­tial outcome (Yijk) for individual k in period j (j= 1, 2) of sequence i (i=1, 2) may be mod­eled as a function of treatment effect (τd[I,j], is treat­ment effect in period j of sequence i), pe­riod effect (Πj , is jth period effect), effect of sub­ject k in sequence i (Sik), carryover ef­fect (λ) and error term (εijk). When there is no carry-over effect and all subjects comply with their as­signed treatment, the widely used model is (10):

Yijk = μ + τd[I,j] + Πj + Sik + εijk

(i=1,2 ;   j=1,2 ;  k=1,2,…,n),                    [1]

The equation (1), standard model is appro­pri­ate to estimate of treatment effects without non-compliance. Since there is non-compli­ance we suggest equation 2 (adjusted model) that outcome was modeled by treat­ment effect (τd[I,j]), period effect (Πj), ef­fect of subject k in sequence i (Sik) and error term (εijk).


where, R=r (r=A, B) to denote the assigned treatment and is indicator of patient at period j of sequence i. We use D(R) to denote the observed treat­ment received of the subject in Jth period of ith sequence with assigned treatment r. D(r)= r (r= A, B) if the subject k took more than 50% of the assigned dose of R=r and D(r)=0 otherwise. We let  Yijk(R,D(R)) to denote the po­ten­tial outcome of the kth subject in the Jth period of the ith sequence with as­signed treatment R and treatment received D(R), which has a normal distribution.  is an indicator for ob­served treatment received for subject k in the period j of the sequence i; Sik is the random ef­fect of the kth subject in the ith sequence, which has a normal dis­tri­bution with mean 0 and variance σ2s; εijk is the random error term, which has a normal dis­tri­bution with mean 0 variance σ2e. In this study, Chi-square or Student's t-Test tests were used to analysis of baseline demographic and scores. Statistical analysis was performed by SAS In­stitute Inc. Version 9.1 (2002). All sta­tisti­cal tests were two-sided and were performed at the 0.05 significance level.


Forty two patients participated in the present study. Thirty (71.4%) were female (Table 2). One third of participants had family history of joint arthritis. Based on clinical symptoms and results of radiography, 6 patients (14%) had low arthritis, 15 patients (36%) had moder­ate ar­thri­tis and 21 patients (50%) had severe ar­thritis. However, the difference was not sta­tis­tically significant in both treatment groups (P>0.05). We did not evaluate the side effect of the new treatment (MM po­made). There was not a sta­tistical significant difference at baseline scores between herbal joint pomade and placebo (Table 2) for pain, physical function, stiffness, and os­teoarthri­tis intensity. There was not carry over ef­fect.

Compliance to assigned drug dosage be­tween participants was divided into two catego­ries (com­pliance= 1, noncompli­ance= 0; Table 1). Table 3 shows mean of out­comes (osteoarthri­tis intension scores) in two periods having two sequences correspond­ing to complete com­pli­ance (stan­dard model) and non-compliance (adjusted model) assumption. The results in­dicated that MM pomade in comparison with pla­cebo had more positive effects on de­creasing the knee pain and symptoms of arthritis dis­ease where the patients did not have a com­plete com­pliance to the treatment (Table 3). In addi­tion, Table 3 showed that t-test statis­tics cor­responding to equation [1] and [2] was 1.96 and 2.01, respectively. Effect size based on standard model (equation 1) and ad­justed model (equation 2) was 0.62 and 0.64, respectively (see Table 3).

Table 2: Baseline demographic and characteristics of patient in both treatment groups

Placebo (n=21)

MM pomade (n=21)


Age (yr)a




Weight (kg)a




Height (cm)a




Children (number) a




BMI (kg/m2) a




Education (illiterate)b




Sex (Female) b




Pain score a




Physical function score a




Stiffness score a




Osteoarthritis intension score*a




a Data are presented as mean ±SD; b Data are presented as percent.

* Osteoarthritis intension score was a compound of pain, physical function and stiffness scores by principle compo­nent analysis, ranged 0 (no osteoarthritis) to 100 (extreme osteoarthritis)

Table 3: Summary statistics of parameters, without pretreatment variables, under the model (1), based on com­pletely compliance assumption; and model (2), based on non-compliance assumption (Standard deviation in paren­theses)

Period I

Period II

Based on model (1)

Based on model (2)

Mean (SD)

Mean (SD)

Standard Model


Adjusted Model

Standard Model


Adjusted Model






Effect size


Effect size

Sequence BA
















- 1.97*



Sequence AB













* P< 0.01;

SD= Standard deviation; Standard model is a model based on complete compliance assumption or equation (1); Adjusted model is a model based on non-compliance assumption or equation (2); B to denote placebo and A to denote Marhame-Mafasel pomade


Osteoarthritis is the most prevalent chronic non-infective joint arthritis and it does not have an absolute remedy (1). The oral and in­jec­tion forms of existing treatments have sys­temic side effects and are not recom­mended for a long time. MM pomade does not have sys­temic side effects. MM pomade like Piroxicam gel, Diclofenac ointment, and comfrey root ex­tract ointment (11-20) has suit­able anti-inflam­mation effect. Cap­saeicine ointment (chili ex­tract) has cutane­ous and mucoid side effects, while MM po­made is a suitable pomade with no side ef­fects (like itch, bleb) (21, 22). The palliative ef­fects of MM pomade on painful os­teoarthri­tis of the knee was more than Cop­per-Salicylate gel ointment, because previ­ous studies indicated that efficacy of Cop­per-Sali­cylate gel ointment was similar to placebo ef­fects; and, occasionally it had se­vere side effect (19). This study like other randomized trials may be marred by devia­tions from protocol, notably some patients failing to comply with the prescribed treat­ment. Therefore, we adjusted the treatment ef­fects (τD) corresponding to com­pliance lev­els.

In this study, we estimated conditional aver­ages and variances in two periods including two sequences (based on equation 2 or ad­justed model) rather than unconditional sum­mary sta­tistics (based on equation 1 or stan­dard model).

In conclusion, herbal joint pomade "Mar­hame-Mafasel" in comparison with placebo has more positive analgesic effects on pri­mary knee os­teoarthritis. In addition, accord­ing to these find­ings, treatment effect should be adjusted for non-compliance in random­ized trials.


The EHJP study was supported by Shahed Uni­versity. Pomade "MM" was made by Phar­macology Division of Shahed Univer­sity, Iran; which its formula exists at Pharma­cology Di­vision of Shahed Univer­sity. We thank all trial site investigators for their dedication on data collected, registry and patients follow up.  The authors declare that they have no conflicts of in­terest.


  1. Parmet S, Lynm C, Glass RM. Os­teo­arthritis of the knee. J Am Med Assoc. 2003; 289(8):1068.
  2. Peat G, McCarney R, Croft P. Knee pain and osteoarthritis in older adults: a re­view of community burden and cur­rent use of primary health care. Ann Rheum Disorder. 2001; 60: 91-7.
  3. Felson DT. Osteoarthritis of the knee. N Eng J Med. 2006; 354: 841-8.
  4. Felson DT. Epidemiology of os­teo­arthritis. 2nd ed. Oxford, UK, Ox­ford Uni­versity Press, England, 2003: pp.9-16.
  5. Dillon CF, Rasch EK, Gu Q, Hirsch R. Prevalence of knee osteoarthritis in the United States: arthritis data from the Third National Health and Nutrition Ex­ami­nation Survey 1991-1994. J Rheum. 2006; 33(11):2271-9.
  6. Guccione AA, Felson DT, Anderson JJ. The effects of specific medical con­di­tions on the functional limitations of eld­ers in the Framingham study. Am J Pub­lic Health. 1994; 84:351-8.
  7. Katzung BG. Basic and clinical phar­macology. 8th ed. McGraw - Hill Inc, Canada, 2001: pp.1088-1089.
  8. Cushnaghan J, Diappe PA. Study of 500 patients with limb joint osteoarthri­tis: Analy­sis by age, sex and distribution of symptomatic joint sites. Ann Rheum Dis­order. 1991; 50: 8-13.
  9. Mehdibarzi D, Naseri M, Faghihzadeh S, Kamalinejad M, Bahrami M. Effi­cacy of Herbal joint pomade“A” in the treatment of primary osteoarthritis of the knee: a randomized, double blind clinical trial. Daneshvar Medicine Jour­nal. 2008; 16 (77): 49-56.
  10. Jones B, Kenward MG. Design and analysis of cross-over trials. 2nd ed. Chap­man & Hall Inc. USA, 2003: pp.25-49. 
  11. Trnavsky K, Fisher M, Vogtle-Junkert U, Schreyger F. Efficacy and safety of 5% ibuprofen cream treatment in knee osteoarthritis. J Rheum. 2004; 31:565-72.
  12. Van Haselen RA, Fisher PA. A ran­domized Controlled trial Comparing topi­cal piroxicam gel with a homeo­pathic gel in osteoarthritis of the knee. Rheu­ma­tology. 2000; 39(7): 714-9.
  13. Hatori M, Kokubun S. Clinical evalua­tion of indomethacin-containing patches for osteoarthritis and extremity trauma. Curr Med Res Opin. 1997; 13(9): 511-5.
  14. Ottillinger B, Gomor B, Michel BA, Pavelka K, Beck W, Elsasser U. Effi­cacy and safety of eltenac gel in the treatment of knee osteoarthritis. Os­teo­arthritis Cartilage. 2001; 9(3): 273-80.
  15. Biswal S, Medhi B, Pandhi P. Long term efficacy of topical non-steroidal ant-inflammatory drugs in knee os­teo­arthritis meta-analysis of random­ized pla­cebo controlled clinical trials. J Rheum. 2006; 33(9): 1841-4.
  16. Banning M. The use of topical di­clofenac for pain in osteoarthritis of the knee: a review. Br J Commu Nurse. 2006; 11(11): 487-92.
  17. Baer PA, Thomas LM, Shainhouse Z. Treatment of osteoarthritis of the knee with a topical diclofenac solution: a ran­domized controlled 6 week trail. BMC Musculoskeletal Disorder. 2005; 6: 44-48.
  18. Machen J, White field M. Efficacy of a proprietary ibuprofen gel in soft tissue injuries: a randomized, double blind, pla­cebo-controlled study. Inter J Clinic Prac. 2002; 56(2): 102-6.
  19. Shackel NA, Day RO, Kellett B, Brooks PM. Copper-Salicylate gel for pain relief in osteoarthritis: a random­ized controlled trial. Med J Aust. 1997; 167(3): 134-6.
  20. Grube B, Grunwald J, Krug L, Staiger C. Efficacy of a comfrey root extracts ointment in the treatment of patients with painful osteoarthritis of the knee. Phy­tomedicine. 2007; 14(1): 2-10.
  21. Schnitzer, Thomas J. Non NSAID phar­macological treatment options for the man- ­agement of chronic pain. AmJj Med. 1998; 105(1B): 458-523.
  22. Rosenstein, Elliot D. Topical agents in the treatment of the Rheumatic disor­ders. Rheum Disorder Clinic North Am. 1999; 25(4): 899-918.

JRHS Office:

School of Public Health, Hamadan University of Medical Sciences, Shaheed Fahmideh Ave. Hamadan, Islamic Republic of Iran

Postal code: 6517838695, PO box: 65175-4171

Tel: +98 81 38380292, Fax: +98 81 38380509

E-mail: jrhs@umsha.ac.ir