Version Nov 16 Nov 2010 RK/SW 1
World Workshop on Oral Medicine V
Management Issues in Oral Submucous Fibrosis: a Systematic Review of Medical
Interventions and Recommendations for future research.
Saman Warnakulasuriya (Section Head)
Ross Kerr (Reviewer)
Alan Mighell (Reviewer)
Thomas Dietrich (Reviewer)
Ajura Jalil (Assistant Reviewer)
Jyotsna Rimal (Assistant Reviewer)
Mona Nasser (Consultant)
Michael Bornstein (Consultant)
Farida Fortune (Consultant)
Vinay Hazarey (Consultant)
Newell Johnson (Consultant)
Toru Nagao (Consultant)
Peter Reichart (Consultant)
Sol Silverman (Consultant)
Short title: Oral submucous fibrosis, medical management
Key words: Oral submucous fibrosis, systematic review, management, research
Notes from Saman to Reviewers & Consultants: please would you add your affiliation as
brief as possible here? To be incorporated by us to title. After the first 4 names (section
head and 3 consultants) I propose to arrange other co-authors listed alphabetically.
Version Nov 16 Nov 2010 RK/SW 2
Oral submucous fibrosis (OSF) is a chronic, insidious disease that affects the lamina propria of
the oral mucosa and, as the disease advances, it involves tissues deeper in the submucosa of the
oral cavity with resulting loss of fibroelasticity. The disease manifests with blanching and
stiffening of the oral mucosa leading to limitation in opening of the mouth. The presence of
fibrous bands in lips, cheeks and soft palate is a hallmark of the disease. Oral submucous fibrosis
predominantly affects South, South Asian and East Asian populations and is seen in India,
Pakistan, Bangladesh, Nepal, Sri Lanka, southern parts of China, Taiwan, Melanesia and
Micronesia and in the pacific islands. The disease is also reported among Asian migrant
communities living in the South and East Africa, parts of Europe, and in North America.
OSF was first described by Schwartz in 1952 among five Indian females living in Kenya and
he coined the term Atrophia idiopathica (trophica) mucosae oris. Several other descriptive terms
have been attributed, submucosal fibrosis of palate and pillars, diffuse oral submucous fibrosis,
idiopathic scleroderma of the mouth, idiopathic palatal fibrosis, and sclerosing stomatitis. The
etiology of the disease was thought to be multifactorial and several agents have been reported,
including local irritants (chillies), nutritional deficiency, genetic predisposition, and auto-
immune disease. There is now conclusive evidence that OSF is caused by areca nut, a
masticatory substance used by Asians (Gupta & Warnakulasuriya, 2002; IARC, 2004). Several
mechanisms and biological pathways have been proposed for the pathogenesis of the disorder, all
based on the constituents of areca nut and genetic susceptibility to the disease (Tilakaratne et al.,
2006, Rajalalitha & Vali, 2005). In essence the disease could be described as a collagen
metabolic disorder with changes observed in the extracellular matrix of the lamina propria and
Version Nov 16 Nov 2010 RK/SW 3
submucosa of the oral cavity due to both increased collagen synthesis and/ or reduced collagen
degradation. Areca nut is the fourth most addictive substance in the world (Gupta &
Warnakulasuriya, 2002), a dependency syndrome has been described (Winstock, 2002) and
unlike for tobacco cessation, interventional programmes have neither been adopted nor evaluated.
There is a significant variation in the prevalence of OSF in different countries and communities
directly attributable to the patterns of areca nut use, age of onset of the habit and regional
variations in the available product. Regional variations in the prevalence of the disease has been
observed in India. Reports on the frequency of encountering OSF suggests that the disease has
rapidly risen in India from an estimated 250,000 in 1980 to 2 million cases in 1993. The reasons
for the rapid increase of the disease is hypothesised due to commercial marketing strategies of
the pan masala industry that produces and markets freeze-dried preparations of areca nut, and an
increased uptake of this habit by young people (Gupta et al., 1998). The disease has even been
identified in infancy, since a 4 year old child was reported to have developed OSF in Canada
(Hayes, 1985). At the same time the pattern of use of areca nut has also changed in other parts of
south Asia. For example, in Thailand and Cambodia the use of areca nut has been decreasing for
Malignant potential in OSF was described by Pindborg and Sirsat in 1966 (Pindborg & Sirsat,
1966). In a long term follow up study the annual transformation rate was approximately 0.5%
(Murti et al., 1985). OSF is now a well recognised potentially malignant disorder of the oral
cavity (Warnakulasuriya et al., 2007). Various classification systems for OSF based on clinical
and histopathological criteria were reviewed by Ranganathan and Mishra (2005). There are no
established markers to identify who may be predisposed to the disease, nor to identify the risk of
malignancy in affected individuals.
Version Nov 16 Nov 2010 RK/SW 4
The treatment of OSF has been reviewed previously, including a narrative review by Jiang
and Hu (Jiang & Hu, 2009) and a Cochrane review by Fedorowicz et al (Fedorowicz et al., 2008).
The review by Jiang and Hu dealt with ‘the role of drugs’ in the treatment of OSF. The review
included a total of 15 publications (involving 1,224 patients), six of which were classed as
randomized controlled trials, four as controlled clinical trials, and five as ‘other experimental
studies’. However, it is unclear how these papers were selected for inclusion in the review.
Overall, the authors concluded that the ‘effect’ of various drug treatments was ‘not satisfactory’
and that the research in this field was insufficient.
The Cochrane review by Fedorowicz et al. had the objective to ‘assess the effectiveness of
interventions in the management of pain and restricted jaw opening or movement occurring as a
result of oral submucous fibrosis’. Only randomized controlled clinical trials of patients with
trismus or restricted jaw movement and a confirmed diagnosis of OSF (by clinical examination
and biopsy) were considered. Pre-specified primary outcomes included (i) resumption of normal
eating, chewing and speech, (ii) change or improvement in maximal jaw opening, measured as
interincisal distance, (iii) improvement in range of jaw movement utilizing any validated
assessment tool, and (iv) change in severity of oral/mucosal burning pain using any recognized
validated pain scale. Secondary outcome measures included (i) postoperative discomfort or pain
as a result of the intervention: patient assessed using any validated pain scale, (ii) length of
hospital admission, (iii) quality of life (QOL) as assessed by any validated questionnaire, either
generic or oral health specific, and (iv) patient satisfaction assessed by validated questionnaire.
In addition, healthcare costs and adverse effects were to be considered. After review of
potentially eligible studies, 2 studies involving 87 participants were included. The validity of
both of the included studies was rated as having a high risk of bias, i.e., plausible bias that
Version Nov 16 Nov 2010 RK/SW 5
seriously weakens confidence in the results. In terms of results regarding the primary outcome
measures, both trials included measurements of change of interincisal distance; however
incomplete reporting of results hampered the ability of reviewers to draw quantitative
conclusions or to corroborate the reported scores. No data on resumption of normal eating,
chewing and speech, or range of jaw movement were reported in either of the included studies.
Changes in severity of oral/mucosal burning pain were not assessed using validated pain scales
and data were considered of insufficient quality to draw any conclusions. No data regarding any
of the secondary outcome measures or costs, and no quantitative data regarding adverse events
were reported in either of the studies. The authors concluded that the uncertain validity of a
limited amount of available data would not appear to support the view that any of the evaluated
interventions were effective, beneficial or safe. The authors also highlighted several issues
regarding the design and reporting of future trials, including recommendations for stratified
randomization or minimization for treatment allocation based on baseline disease severity, for
rigorous blinding and improved methods of outcome assessments and the use of validated
instruments to ascertain relevant outcomes, as well as reporting of trials in accordance with
CONSORT standards (Moher et al., 2001). However, the authors also acknowledge the
challenges and difficulties faced by investigators in low and middle income countries in which
OSF is prevalent.
Study of the natural history of OSF show that it is an insidious disorder that progresses with
time. In clinical practice there are a number of treatments for OSF, ranging from medical
interventions, surgical interventions, physical therapy, and of course habit control (ie cessation of
areca nut use). Often a combination of strategies is used.
Version Nov 16 Nov 2010 RK/SW 6
In general, chronic chewers with OSF seem to complain of two problems: inability to open
their mouths and function normally, and a burning sensation and intolerance to spicy foods that
are often the mainstay of the diet, leaving an individual handicapped both physically and
psychologically. The severity and blend of signs and symptoms of OSF is highly variable.
Patients with mild early disease, marked by a strong inflammatory component, are less likely to
have fibrosis and more likely to complain of burning. This is in marked contrast to those with
severe advanced disease where irreversible fibrosis and loss in function predominates. As such
the aims of treatment are to reverse or ameliorate these signs and symptoms, and in addition, to
minimize the risk for malignant transformation. There is a dizzying array of reported medical
interventions including dietary supplementation (vitamins, anti-oxidants), anti-inflammatory
agents (principally corticosteroids) and proteolytic agents (such as hyaluronidase and placental
extracts) and anti-cytokines. Such agents may be administered orally, topically or via
submucosal injection. Surgical interventions are generally reserved for more advanced cases of
OSF. Physical therapy may be used as a single modality intervention or combined with other
interventions (principally surgical interventions).
Our objectives are:
a. To develop a systematic map on the current medical (ie non-surgical) interventions
available for the management of OSF.
b. To update the evidence on the medical interventions for the management of OSF.
Version Nov 16 Nov 2010 RK/SW 7
c. To develop a structure implication for a low-cost research protocol for future clinical
trials in this field, with an emphasis on conducting studies in regions of the world
where OSF is prevalent.
3. Review Methodology
Detailed automated searches of PubMed were conducted using “oral submucous fibrosis” as the
key words up to September 2010. Interventional studies were then selected from the abstracts.
Additional searches of the Indian and Chinese literature were manually conducted. Chinese
studies of interest were translated into English. The initial pool of intervention studies and
review articles were searched for references leading to additional papers missed in the automated
searches. Articles that were case reports and statements of expert opinion were only included if
they offered some possible insight.
Interventional studies were then categorized by study type, including randomized controlled
studies (RCTs), observational studies, or case series reports. To meet the criteria for RCTs, the
study had to be prospective, include a control group and state that subjects were randomly
assigned to the control and interventional groups. Observational studies included uncontrolled
(or poorly controlled) and/or non-randomized prospective study of a single intervention OR
retrospective studies comparing two or more different interventions. A case series constituted a
retrospective series of cases based on a single intervention. Participants included individuals in
any age group with a confirmed diagnosis, by clinical examination and/or biopsy of OSF. Types
Version Nov 16 Nov 2010 RK/SW 8
of interventions included habit intervention, surgical procedures, medical treatments (ie systemic,
submucosal injection or topical agents, or physical therapy. The primary outcomes explored
were the (i) objective change or improvement in maximal jaw opening, measured as the inter-
incisal distance, (ii) subjective change in severity of oral/mucosal burning pain using any
recognised validated pain scale, (iii) subjective change in quality of life using any questionnaire,
whether validated or not, and (iv) reduction in the rate of malignant transformation.
The secondary outcomes explored were any other objective or subjective changes, such as
adverse events, and improvement of anaemia and co-morbidities.
Data collection and extraction:
Studies selected were evaluated independently by three reviewers (RK, TD & AM), and a data
extraction table was developed for this purpose. For each study the following data was captured:
study period, publication language, country, study setting, number of subjects, study type,
intervention types, design details (ie control group, randomisation, blinding, timing of visits),
description of population (gender, age, diagnostic criteria for OSF, baseline disease severity,
habit profile), outcomes measured (subjective and objective), follow-up information, and details
about statistical analyses.
The quality of evidence for each study was assessed using the Risk of Bias table (Higgins, Green,
Version Nov 16 Nov 2010 RK/SW 9
Due to clinical heterogeneity of the studies and missing data, we were not able to pool the data of
the included RCTs and provided a narrative synthesis of the data
GRADING the overall quality of evidence:
The overall quality of evidence for each outcome was assessed and reported using the GRADE
approach (Guyatt et al., 2008a), (Guyatt et al., 2008b), (Guyatt et al., 2008c), (Guyatt et al.,
2008d), (Jaeschke et al., 2008).
The quality of evidence for the questions were presented and discussed in the consensus group.
The balance between risk and benefits, necessary cost and resources and patients values and local
contexts has been taken into consideration. The final recommendations were graded from strong
or weak based on the judgment of all participating experts.
50 publications were included in the pool of investigations on the treatment of OSF, of which 23
were surgical in focus (and will be reported elsewhere by our group). Of the remaining studies,
22 were medical, 3 were medical/surgical, one was medical/physical, and one was
medical/surgical/physical. The earliest study was reported in 1980 from India, and approximately
half were undertaken after the year 2000. Three were reported in Chinese and the rest in English.
Of these 27 studies, about half were conducted in India, about a third in Taiwan/China, a small
number in other South Asian countries, and two among immigrants living abroad (one each in
the UK and USA). All of the reported investigations were hospital/institution based and none
were conducted in community settings.
Version Nov 16 Nov 2010 RK/SW 10
In most of the 27 investigations, the diagnosis of OSF was based on the classic clinical
presentation. Histopathology was used to confirm diagnosis in 12 of the investigations. Enrolled
populations had a wide spectrum of OSF (ie from early to advanced), and yet stratification of the
study group by OSF stage/severity rating was defined at baseline in only approximately half of
the studies, most of these by reporting baseline opening, although some studies grouped subjects
by range of opening (Ariyawardana et al., 2005), (Maher et al., 1997) and (Lai et al., 1995), or by
using various rating scales (Khanna & Andrade, 1995), (Gupta et al., 1992), (Talsania et al.,
2009) and (Singh et al., 2010).
Baseline demographic information, such as age, gender was reported in 60% of studies.
Baseline patterns of areca nut use were reported in 30% of investigations, and those of alcohol
and/or tobacco were reported in 22%. Baseline assessment of nutritional or dietary habits was
reported in a single study (Tai et al., 2001), and laboratory assessment of hematologic status was
made in 30% of studies.
A summary description of the study types is set out in Table 1. Only four studies
(Rajendran et al., 2006), (Kumar et al., 2007), (Cox & Zoellner, 2009) and (Jirge et al., 2008)
met our criteria for an RCT (including the two (Rajendran et al., 2006) and (Kumar et al., 2007)
previously reported in the Cochrane review), and all were run at a single center. There was one
other prospective controlled study that lacked randomization (Lin & Lin, 2007). The rest were
rated as observational or retrospective studies.
Tables 2 and 3 highlight the different interventions and how they are alleged to work.
There were no studies that looked at the effect of habit control alone as the primary endpoint, ie
cessation of areca nut habits. The methodology of 14 studies included the advice to quit the habit,
although only two of these described specific measures for cessation. In these studies subjects
Version Nov 16 Nov 2010 RK/SW 11
were given a dental cleaning at baseline to remove staining and then re-examined at follow-up
visits for any new staining (Ariyawardana et al., 2005) and (Kumar et al., 2007). No serum
markers for metabolites of areca nut were utilized.
15 studies used a single agent, and the rest studied combinations of agents. 22 studies
included the use of nutrients, micronutrients and/or anti-oxidants, 21 studies included the use of
immunomodulatory agents that reduced the inflammatory component, principally injected
corticosteroids (16 studies). 19 studies included the use of proteolytic enzymes to reduce fibrosis
of which 7 used hyaluronidase. 4 studies included agents to promote blood flow. Agents were
delivered orally for systemic absorption, intra-lesionally, or topically.
Outcome measures reported in these studies were highly variable both in the type and the
manner in which they were measured. In terms of objective measures, mouth opening (generally
measured as inter-incisal opening) was the most consistently and reliably measured outcome
across all studies. Other objective measures included changes in tongue movement (ie ability to
protrude), degree of suppleness of the tissues, amount of blanching of the mucosa, presence of
ulceration/vesicle formation, and amount of dorsal tongue papillation, although the methodology
for measuring these other objective outcomes was poorly defined and of questionable
unreliability. In terms of subjective measures, oral burning/pain was the most consistently
measured subjective outcome, although very few studies reported using validated pain
assessment instruments, such as a visual analogue pain rating scale. Other subjective measures
included change in taste, oral dryness, and ability to chew, swallow, or speak. None of the
studies used validated instruments evaluating quality of life of subjects with OSF, nor could we
find any such instruments in the published literature.
Version Nov 16 Nov 2010 RK/SW 12
22 investigations did not specify whether or not subjects completed a given treatment regimen.
Of the remaining studies, 17 reported >75% of the subjects completed the study regimen.
Follow-up of subjects after treatment was highly variable, with only 19 studies reporting follow-
up beyond 1 year.
Randomized Controlled Studies:
There were three RCTs describing investigation of medical interventions: pentoxifylline
(Rajendran et al., 2006), lycopene (Kumar et al., 2007) and levamisole with anti-oxidants (Jirge
et al., 2008). Rajendran et al. divided the 29 participants into two groups that took either oral
pentoxifylline or multi-vitamins. All those enrolled completed the 7 month study period. The
authors reported statistically significant improvements in the oral pentoxifylline group (n=14)
compared to controls with respect to objective criteria (mouth opening, tongue protrusion and
relief from circum-oral fibrotic bands) and subjective criteria (intolerance to spices, burning
sensations, tinnitus, difficulty in swallowing, and difficulty in speech).
Kumar et al. recruited 83 participants who were divided between study groups that
received either oral lycopene (n=21; group A), oral lycopene with intralesional corticosteroids
(n=19; group B) or an oral placebo (n=18; group C). The two-month intervention period was
completed by 58 people. Objective measurement of mouth opening was reported to be
significantly improved with an average increase of 3.4mm, 4.6mm and 0mm for groups A, B and
C, respectively. The increases were maintained at 3 and 6-months review. All patients who took
lycopene reported relief from burning sensations within 2 weeks, whereas only one patient from
the placebo group reported a similar improvement.
Version Nov 16 Nov 2010 RK/SW 13
The 45 participants reported by Jirge et al. were divided equally between three study
groups: oral levamisole (group I), an oral antioxidant (group II), or oral levamisole with
antioxidant (group III). On conclusion of the intervention period (approximately 15 weeks) there
was improvement of mouth opening of 7.1%, 6.7% and 8.0% in groups I, II and III, respectively.
These gains were maintained on further evaluation two months later. There was also a significant
reduction in burning sensations in all study groups.
Cox and Zoellner enrolled 54 Nepali subjects into 3 groups: physiotherapy, injections
with combination hyaluronidase/steroids, and a control group. After 4 months, subjective and
objective measures were compared to baseline. The physiotherapy group showed a significant
increase in opening but had no superior effect on subjective measures.
Re-analysis by the working group of the published data from these four RCTs using the
GRADE criteria identified significant limitations with each report and challenged the
conclusions reached by the authors (Table 4) .
5. Future Studies
The review team, echoing the sentiments of other reviewers, appreciates the opportunity and
importance to offer suggestions and recommendations for future research. Clinical research
methodology has evolved rapidly in the industrialized world, yet not all parts of the world have
the experience, nor the necessary infrastructure, to design let alone run randomized controlled
trials. While the methodological issues in the published literature we reviewed offer weak
evidence at best to make recommendations for the management of patients with OSF, there is
much valuable insight to be gained from the studies we reviewed.
Version Nov 16 Nov 2010 RK/SW 14
Moving away from the perspective of a systematic review, and focusing on mining the
studies for information to help direct future research, we developed a list of objectives (adapted
from (Brown et al., 2006)), and summarized in Table 6.
a. What populations should be researched?
b. What types of interventions are needed?
c. What types of outcomes should be measured?
d. What study designs are needed?
e. What infrastructure is required to conduct studies?
The populations for research on OSF are dictated by where the areca nut habit is prevalent.
Studies must be conducted in both in South East Asia (India, Pakistan, Nepal, Bangladesh, Sri
Lanka), and in Chinese populations (Taiwan and Southern China) where studies have already
been performed and the research infrastructure is developing. Numerous other countries have
high rates of areca nut use (Burma, parts of Malaysia, Pacific islands and others) although very
little literature suggests that clinical studies are ongoing. Additional studies conducted in
immigrant populations, such as the UK, US, or Australasia, are secondary, yet have the potential
to overcome some of the methodological flaws inherent in countries where clinical research
infrastructure is lacking.
Given the variable spectrum of the signs and symptoms of OSF, subpopulations of
patients grouped by disease severity/stage should be studied separately because different
interventions may be effective at different stages of pathogenesis. For simplicity, there are two
distinct populations: those with advanced-stage disease hallmarked by irreversible and
debilitating fibrosis, and those who have not reached advanced-stage. Studies must define
Version Nov 16 Nov 2010 RK/SW 15 Download full-text
specific inclusion and exclusion criteria to foster the enrollment of subjects suited to the type of
intervention. In terms of demographics, studies are needed not only in adult populations, but also
in children who are regularly using areca nut products (particularly gutkha) (Gupta & Ray, 2003).
There may also be differences in OSF populations related to the habits and types of areca-nut
We hypothesize that habit cessation alone as an intervention may have a large effect, more so on
the symptoms of OSF rather than reversing fibrosis. The almost complete lack of studies
incorporating habit control suggests that investigators have difficulty managing the dependence
on areca nut products. Indeed the introduction of gutkha into the marketplace in India has led to
even higher rates of dependence and OSF (Gupta, 1998). Future interventions must incorporate a
standardized preventive plan even if a high relapse rate is anticipated, and include methodology
to allow investigators to control for relapse during the study and follow-up during data analysis.
Serial measures of serum areca alkaloids might be the gold standard to detect relapse or
continued use of areca products during the studies, although simple strategies such as performing
a baseline dental prophylaxis to remove extrinsic staining and re-evaluate for new staining might
be an excellent surrogate.
Our current understanding of the pathogenesis of OSF (Tilakaratne et al., 2006) includes
overlapping phases, an early inflammatory phase and the later fibrosis phase, suggesting that
interventions can be tailored to the severity of disease. At one end of the spectrum, new studies
for the treatment of advanced disease are needed. We know that surgical excision of fibrosis will
provide short-term improvement in function. Yet, there are a number research questions