This clinical trial evaluated the effectiveness of prolonged treatment with a pentoxifylline-tocopherol-clodronate combination (PENTOCLO) for refractory mandibular osteoradionecrosis (ORN) in 54 patients. PENTOCLO treatment for a mean of 16 months resulted in: 1) A significant and progressive reduction in exposed bone, with complete mucosal and bone healing in all patients. 2) Significant improvement in pain and symptoms. 3) No safety issues were observed. 4) PENTOCLO appears to be an effective and safe treatment for refractory ORN, though a randomized controlled trial is still needed.

1. Int. J. Radiation Oncology Biol. Phys., Vol. 80, No. 3, pp. 832–839, 2011
Copyright Ó 2011 Elsevier Inc.
CME
Printed in the USA. All rights reserved
0360-3016/$–see front matter
doi:10.1016/j.ijrobp.2010.03.029
CLINICAL INVESTIGATION Normal Tissue
COMPLETE RESTORATION OF REFRACTORY MANDIBULAR
OSTEORADIONECROSIS BY PROLONGED TREATMENT WITH
A PENTOXIFYLLINE-TOCOPHEROL-CLODRONATE COMBINATION (PENTOCLO):
A PHASE II TRIAL
SYLVIE DELANIAN, M.D., PHD.,* CECILE CHATEL, D.D.,y RAPHAEL PORCHER, PH.D.,z JOEL DEPONDT, M.D.,
´
PH.D.,x AND JEAN-LOUIS LEFAIX, PH.D.{
*Service d’Oncologie-Radiotherapie, and zDepartement de Biostatistique et Informatique Medicale, Hopital Saint-Louis, APHP, Paris;
´ ´ ´ ˆ
y
Odontologie, Institut Gustave Roussy, Villejuif; xService de Chirurgie Cervico-Faciale, Hopital Bichat, APHP, Paris; and {CEA/DSV/
ˆ
IRCM-LARIA, CIRIL-GANIL, Caen, France
Purpose: Osteoradionecrosis (ORN) is a nonhealing wound of the bone that is difficult to manage. Combined treat-
ment with pentoxifylline and vitamin E reduces radiation-induced fibrosis and ORN with a good prognosis. We
previously showed that the combination of pentoxifylline and vitamin E with clodronate (PENTOCLO) is useful
in healing sternocostal and some mandibular ORN. Is PENTOCLO effective in ORN of poor prognosis?
Methods: 54 eligible patients previously irradiated for head and neck cancer (among 72 treated) a mean 5 years
previously received exteriorized refractory mandibular ORN for 1.4 ± 1.8 years, mainly after local surgery and
hyperbaric oxygen had been ineffective. The mean length of exposed bone (D) was 17 ± 8 mm as primary endpoint,
and the mean Subjective, Objective, Management, and Analytic evaluation of injury (SOMA) score was 16 ± 4.
Between August 2000 and August 2008, all patients were given daily oral PENTOCLO: 800 mg pentoxifylline,
1,000 IU vitamin E, and 1,600 mg clodronate 5 days per week alternating with 20 mg prednisone and 1,000 mg
ciprofloxacin 2 days per week. The duration of treatment was related to consolidated healing.
Results: Prolonged treatment (16 ± 9 months) was safe and well tolerated. All patients improved, with an exponen-
tial progressive—(f[t] = a.exp(-b.t)—and significant (p < 0.0001) reduction of exposed bone (D), respectively
(months): D2 À42%, D4 À62%, D6 À77%, D12 À92%, and D18 À96%, combined with iterative spontaneous seques-
trectomies in 36 patients. All patients experienced complete recovery in a median of 9 months. Clinical improve-
ment was measured in terms of discontinuation of analgesics, new fracture, closed skin fistulae, and delayed
radiologic improvement: SOMA6 À64%, SOMA12 À89%, and SOMA30 À96%.
Conclusion: Long-term PENTOCLO treatment is effective, safe, and curative for refractory ORN and induces mu-
cosal and bone healing with significant symptom improvement. These findings will need to be confirmed in a ran-
domized trial. Ó 2011 Elsevier Inc.
Pentoxifylline, Alpha tocopherol, Clodronate, Osteoradionecrosis, Radiotherapy.
INTRODUCTION tal extraction, mostly 6 months to 5 years after irradiation (1).
Mandibular ORN symptoms, excluding tumor recurrence, are
Mandibular osteoradionecrosis (ORN) is a delayed injury
diverse, ranging from occult disease to major bone destruction
caused by failure of bone healing several years after head-
with soft tissue necrosis and spontaneous complications
and-neck cancer irradiation. Severe ORN can be life-
like osteomyelitis, fistulation, and fracture (2). Multiple risk
threatening and compromise functional prognosis. Although
factors predispose to its development: treatment-dependent
conformal radiotherapy (RT), by improving the therapeutic ra-
factors, including radiotherapy (dose, volume, brachyther-
tio, has reduced the incidence of severe complications, ORN is
apy), surgery (number, volume, hematoma, infection), and
still unavoidable in a mean 10% of cases, especially after den-
Note—An online CME test for this article can be taken at http:// Supported by the Delegation a la Recherche Clinique of the As-
´´ `
astro.org/MOC. sistance Publique des Hopitaux de Paris.
ˆ
Reprint requests to: Dr. Sylvie Delanian, M.D., Ph.D., Service Conflict of interest: none.
d’Oncologie-Radiotherapie, Hopital Saint Louis, 1 Ave Claude Vel-
´ ˆ Acknowledgment—The authors thank Charles Guedon and physi-
lefaux, 75010 Paris, France. Tel: (33) 1-42-49-97-89; Fax: (33) 1- cians from several Parisian institutions for entrusting their patients
42-49-91-97; E-mail: sylvie.delanian@sls.ap-hop-paris.fr to us for treatment.
Presented at the 18th Congress of Societe Francaise de Radiother-
´´ ¸ ´ Received Oct 2, 2009, and in revised form Feb 8, 2010. Accepted
apie Oncologique, Paris, November 2007, and the Special Work- for publication March 10, 2010.
shop of The Royal College of Surgeons of England, London,
November 2007.
832

2. PENTOCLO trial in osteoradionecrosis d S. DELANIAN et al. 833
concomitant chemotherapy; and patient-dependent factors, in- that was not measurable because of trismus, nonexposed bone, or
cluding age, hypersensitivity, high blood pressure, diabetes, maxillary lesion, and 9 withdrew because of a change of mind before
and collagen vascular diseases (3). However, the increased in- the study (2 patients), concomitant progressive cancer (4 patients),
cidence and severity of ORN are mainly due to poor dental sta- fatal sepsis (1 patient), human immunodeficiency virus disease (1 pa-
tient), and intolerance combined with lupus (1 patient). Conse-
tus, local biopsy sites, bone proximity to the initial anatomic
quently, 54 eligible treated patients had complete, available, and
tumor site, or excessive tobacco and alcohol consumption (1). long-term homogeneous data (Table 1). Informed consent was ob-
The management of ORN is usually based on conventional tained from all patients, and the study was approved by the Hospital
medical care focusing on comorbidity factors such as optimi- Ethics Committee. The patients (43 men, 79%), all of whom had
zation of oral hygiene (alcohol and tobacco consumption, a history of tobacco and alcohol consumption, showed no evidence
mouth baths), infection control (antibiotics), and devitalized of recurrent disease on entering the trial.
tissue removal (sequestrectomy) for ORN with a good progno-
sis (1, 3). Furthermore, its incidence is reduced by preventing RT damage
The ORN was caused by standard RT of head-and-neck cancer
trauma such as limited dental extraction. If ORN management
(9–10 Gy/week), with a total prescribed dose ranging from 45–65
always involves conservative measures, recovery at 1 year is Gy (15 postoperative) to 70–75 Gy (39 exclusive RT): RT alone
reported in only 8–33% of patients with a good prognosis (4, (n =13), combined with primary or salvage surgery (n =27), and
5). Hyperbaric oxygen (HBO) is reported to be effective as combined with chemotherapy (n =14). The mean latency period be-
an adjunctive treatment for ORN, but the retrospective trials tween the end of RT and the first ORN symptoms was 4.8 Æ 5.1
involved are restricted (18 patients/study), and recovery years (range, 0.2–29).
ranges from 15% to 43% after HBO alone, vs. 18% to 90%
after HBO combined with limited surgery (4–6). In addition, PENTOCLO treatment modalities
a recent randomized trial in 68 ORN patients failed to One month before inclusion, all patients were given 4-week daily
demonstrate that HBO had any beneficial effect, inasmuch oral disinfiltrating treatment with 20 mg prednisone plus 2 g
amoxicillin-clavulanate plus 1 g ciprofloxacin plus 50 mg flucona-
as only 19% in the HBO group recovered vs. 33% in the
zole to allow further PENTOCLO penetration, which improved
placebo group (7). By contrast, refractory ORN (Epstein Stage
pain and purulence symptoms by a mean of 20% in all patients.
III) required, when technically possible, extensive surgical re-
section or reconstruction, such as hemimandibulectomy or
closure of fistulae using myocutaneous flaps (1, 7). The Table 1. Baseline screening of participants with complete
data analyzed
more recent technique of the facial artery musculomucosal
flap seems to be useful in some patients (8). Characteristic n
Today, no universally accepted treatment exists for this
severe condition. Since 1998, we have proposed, in a new No. of patients N = 54
Age (y): mean Æ SD (range) 59 Æ 10 (30–81)
theory for ORN pathogenesis, that bone damage is mainly Head-and-neck tumor
the result of radiation-induced fibrosis (RIF) (2, 3, 9) and Oral cavity 15
have developed a triple-drug therapy to reduce RIF and Oropharynx 31
bone destruction and to stimulate osteogenesis via the antiox- Other 8 (15%)
idant pathway (10, 11). We have published impressive RT dose level
Exclusive 70–80 Gy 37 (70%)
preliminary results using pentoxifylline (PTX) combined Postoperative 45–65 Gy 17
with vitamin E (PE) in 10 ORN patients with a good Combined head-and-neck cancer treatment:
prognosis and PE boosted by clodronate (PENTOCLO) in RT alone 13 (22%)
the first 8 patients with refractory ORN, with complete Surgery + RT 27
Chemotherapy + RT 14
mucosal recovery in most of them at 6 months (12).
ORN trigger factors:
To assess the maximum efficacy and the time to achieve it, None (spontaneous) 26
and the follow-up during PENTOCLO treatment and long af- After dental extraction 28
terward in patients with refractory ORN, we performed a trial Delays
to define the optimal treatment duration until complete heal- RT-ORN symptoms (y) 4.8 Æ 5.1 (0.2–29)
ORN symptoms/ PENTOCLO (y) 1.1 Æ 1.8 (0.1–12)
ing is established. Failure of previous treatments:
Medical alone 23 (42%)
PATIENTS AND METHODS HBO and/or 13
Surgery (sequestrectomy/flap) 25
Population Baseline ORN characteristics
Between August 2000 and August 2008, 107 ORN patients Mean exposed bone (l+w) mm 17 Æ 7.9
recruited from various centers at Saint-Louis Hospital (Paris) were l mm 24 Æ12 (7–60)
treated with PENTOCLO: 80 consecutive patients with head-and- w mm 10 Æ 6 (2–30)
neck cancer and 27 patients with miscellaneous ORN (5 after cervical Mean SOMA score 15.7 Æ 3.6 (8–24)
RT, 15 sternocostal after breast RT, 7 after pelvic RT) not included Abbreviations: SD = standard deviation; RT = radiation therapy;
for this evaluation (9). Twenty-six of the 80 patients were excluded ORN = osteoradionecrosis; PENTOCLO = pentoxifylline, vitamin
because they did not meet protocol requirements: 8 had measurable E, and clodronate; HBO = hyperbaric oxygen; SOMA = subjective,
good prognosis ORN already healed with PE alone (12), 9 had ORN objective, management, analytic evaluation of injury.

3. 834 I. J. Radiation Oncology d Biology d Physics Volume 80, Number 3, 2011
The PENTOCLO dose was based on pharmacokinetic data, on Table 2. Osteoradionecrosis modified SOMA score
clinical use and long-term safety in other diseases, and on several
dose variations determined by trial and error during our 10 years Subjective (G1–G4)
Pain (occasional/minimal, intermittent/tolerable, persistent/
of clinical experience with 900 RIF, ORN, and plexitis patients intense, refractory)
(13). Each patient was given a daily combination of twice-daily Mastication (difficulty with solid, with soft foods, gastrostomy)
400 mg PTX (800 mg/day) plus 500 IU vitamin E (1,000 IU/day) Objective (G1–G4)
and once-daily 1,600 mg/day clodronate from Monday to Friday Bone exposed (minimum ulceration, <2 cm, 2–4 cm, >4 cm/
(5 days/week), alternating with 20 mg prednisone plus 1,000mg ci- fracture)
profloxacin on the weekend (2 days/week). The PTX dose was Trismus (minimum, 1- to 2-cm opening, 0.5–1 cm, <0.5-cm
reduced from 1,200 to 800 mg/day to avoid adverse effects in pa- opening)
tients without vascular disease. The vitamin E dose provided suffi- Management (G1–G4)
cient antioxidant activity and allowed PTX synergy (increased from Pain (occasional, regular nonnarcotic, regular narcotic, surgery)
500 to 1,000mg/day). Clodronate dose reduction from 7 to 5 days/ Bone exposed (mouth bath, antibiotics, debridement/HBO, large
surgery)
week was sufficient to provide antimacrophagic activity without
Analytic (G1–G4)
causing a calcium problem. Prednisone/ciprofloxacin 2 days/week Mandibular X-ray (questionable changes, osteoporosis/mosaic,
was sufficient to provide intermittent acute anti-inflammatory and sequestra, fracture)
antiseptic action.
The duration of treatment was based on observed progressive Abbreviations: SOMA = subjective, objective, management, an-
ORN regression and on the published long-term effects of PE, to alytic evaluation of injury; HBO = hyperbaric oxygen.
avoid a rebound effect (14).
with treatment was analyzed using nonlinear mixed models. Several
competing models were consecutively considered and compared us-
Outcome measures
ing the Bayesian Information Criterion model (16) including deter-
Participants were reviewed by two investigators before, during,
mination of the random effects and the correlation and variance
and after the end of treatment with 3-year follow-up. Quantitative
structure of residuals that gave the best model fitting. These methods
assessment included measurements of the mean dimensions (D) of
predict the ORN dimensions of a patient measured at a given time-
superficial soft tissue necrosis describing exposed bone osteoradio-
point as a function of time: the predicted change for a given patient
necrosis (EB-orn): (length + width)/2. The primary endpoint was
varies around the average prediction for all patients according to
the relative D regression defined as (D at x months – D at inclusion)/
these random components. The effects of age, time since RT, trigger
D at inclusion, correlated with the extent of recovery of EB-orn
factors, previous treatment, and combined head-and-neck cancer
(Table 1).
treatment on the change in ORN dimensions were tested.
Secondary endpoints included modified Subjective, Objective,
The probability of complete EB-orn healing over time was esti-
Management, and Analytic evaluation of injury (SOMA) score, with
mated using a nonparametric method for interval censored data
qualitative and quantitative personal item variations to allow regular
(17). Model-based predictions were additionally obtained consider-
treatment follow-up (Table 2) (15). Assessment by measuring percent-
ing that ORN dimensions lower than 0.5 mm indicated complete
age changes in D and SOMA score was done every month (1 month
healing, both for the patients in the study and for the population,
[M1]), every 2 months (M2) until complete mucosal healing, then every
the latter by using numeric simulations.
3 months. Patients acted as their own controls (paired data). Regular
All tests were two-sided, and a p value under 0.05 was considered
X-ray and dental computed tomography scans were performed.
as significant. All analyses were performed using R.1.7.1 software
(The R Development Core Team, Vienna, Austria; http://r-project.
Osteoradionecrosis org).
ORN developed a mean 5 years after RT, either spontaneously (26/
54) or after trauma such as dental extraction or biopsy (28/54). ORN
gradually worsened despite medical care including amoxicillin (n = RESULTS
17), HBO (n = 13) and/or local surgical procedures as sequestrectomy
or flap (n = 24). None of these treatments had a lasting beneficial effect Adverse events
on ORN, but they sometimes reduced acute inflammation. At base- Acute safety was satisfactory: no patient stopped the treat-
line, ORN was exteriorized without healing for 1.4 Æ 1.8 years (range, ment because of an adverse event. One patient stopped treat-
0.1–12) with mean EB-orn D0 at 17 Æ 8 mm (Table 1). ment at 1 year: misunderstanding of the disease and
All patients had combined symptoms (Table 2) as local pain or treatment-related epigastralgia, instead of ORN improve-
minimal infection, but 36/54 patients (66.6%) had one or more com- ment.
bined severe symptoms such as skin fistula in 16, chronic osteitis in Twelve of 54 patients (22%) experienced minimal adverse
23 (purulence), facial edema in 6, fracture without shifting in 8, and effects, but were included, like the others, in the analysis.
inferior dental neuropathy in 12. All 54 patients had very severe
Grade 1–2 discomfort during the first weeks of treatment
ORN: one third with Epstein Stage II as 18 chronic persistent
was due to nausea-epigastralgia (4 patients), asthenia (2 pa-
ORN without healing over several months or years, and two-
thirds with Epstein Stage III as 36 active progressive ORN including tients), headache (1 patient), vertigo (1 patient), insomnia
fistula, fracture, or osteitis. The mean baseline SOMA0 score was (1 patient); these patients remained in the study after resolu-
15.7 Æ 3.6 (Table 1). tion by transient (2–4 weeks) reduction in PTX dose (400 mg/
day) or symptomatic treatment with omeprazole or heptami-
Statistical analysis nol; 2 patients with gastrostomy experienced problems with
Data are presented as counts for qualitative variables, and mean crushed PTX tablets. In a previous randomized trial, we
(ÆSD) for quantitative variables. Change in ORN dimensions found no significant differences between PE, PTX, vitamin

4. PENTOCLO trial in osteoradionecrosis d S. DELANIAN et al. 835
Table 3. Mean exposed bone–ORN regression, complete mucosal healing with recovery rate (observed, estimated), and SOMA score
regression during PENTOCLO treatment
No. of treated Mean ORN exposed Complete observed Healing estimated SOMA score
Time patients bone mm (mean Æ SD) recovery rate (%) recovery rate (%) (mean Æ SD)
Baseline 54 17.2 Æ 7.9 15.7 Æ 3.6
2 mo 54 10.3 Æ 7.5 3 (5.5%) 5.6% 11.5 Æ 4.1
4 mo 48 7.3 Æ 7.4 10 (21%) 20.2% 8.5 Æ 4.4
6 mo 46 4.4 Æ 6.1 20 (43.5%) 42.4% 5.8 Æ 4.5
9 mo 43 2.9 Æ 5 26 (60.5%) 59.2% 4.5 Æ 4.2
12 mo 39 1.6 Æ 3.3 26 (67%) 64.6% 3.2 Æ 3.0
18 mo 25 0.8 Æ 2.2 21 (84%) 78.8% 2.0 Æ 2.5
24 mo 20 0.8 Æ 2.4 18 (90%) 85.9% 1.4 Æ 2.2
30 mo 16 0 16 (100%) 100% 0.8 Æ 1.2
36 mo 14 0 14 (100%) 100% 0.4 Æ 0.9
Abbreviations: ORN = osteoradionecrosis; SOMA = subjective, objective, management, analytic evaluation of injury; PENTOCLO = pen-
toxifylline, vitamin E, and clodronate; SD = standard deviation.
E, and double placebo groups in terms of tolerability (dis- (p < 0.0001 as compared with a model with residual EB-
comfort) during treatment (18). When diarrhea was present orn). Model-based predictions fitted the observed values
(1 patient), clodronate was reduced to 800 mg/day, but this quite well for each individual patient’s EB-orn regression,
patient remained in the study for analysis. as illustrated in Fig. 2. This model showed significant EB-
Long-term safety was excellent, with no patient stopping orn regression (p < 0.0001). None of the variables tested
treatment because of severe adverse side effects, but PENTO- was found to modify this treatment effect significantly. Me-
CLO was stopped at 6 months because of transient regressive dian time to complete response was an estimated 9 months
oral exostosis (mandible) in 2 patients, in parallel with a good (Fig. 3).
response (2 patients). After discontinuation of drug. After stoppage of PENTO-
CLO, no rebound EB-orn effect was observed over several
Primary analyses months of follow-up.
Exposed bone regression during PENTOCLO treatment.
PENTOCLO was effective over several months resulting in
exponential EB-orn regression until complete healing with Secondary analyses
mucosal recovery. PENTOCLO was stopped before mucosal All patients responded to treatment and symptom severity
recovery in 15 of 54 patients, who were nonetheless assessed diminished exponentially as assessed by the SOMA score
until their last follow-up. Three patients were immediately
lost to follow-up (before M4). Other causes of loss to
follow-up were: 1 vascular stroke after high blood pressure
(M2), 6 fatal sepsis (M2-M18) due to local severe infection
with facial cellulitis, bone fracture, fistula, or pulmonary in-
fection, because of persistence of co-morbidity factors as
high tobacco-alcohol consumption, HIV, severe undernour-
ishment.), and 5 recurrent or second head-neck or lung can-
cer (M2-M18) as usually observed in such a population (4
patients with progressive cancer were excluded just before in-
clusion). The remaining patients had long-term PENTOCLO
for 16 Æ9 months (6-36 months).
Observed values. (Table 3, Fig. 1)- Mean exposed bone
ORN regression was D2 42 Æ27% at 2 months, D4 62
Æ29%, D6 77 Æ28%, D9 86 Æ23%, D12 92 Æ14%, D18 96
Æ13%, D24 96 Æ14%, D30 and D36 100%.
Modeling. From the observed changes in EB-orn dimen-
sions during treatment, several models were postulated and
the best representative model of the time-course of regression
was found to have the following exponential form: f(t) =
a .exp (-b.t), where t represents the time from treatment onset Fig. 1. Regression of exposed bone osteoradionecrosis during treat-
ment with pentoxifylline, vitamin E, and clodronate individual pa-
in months, and a and b correspond respectively to ORN di- tient data (gray solid lines), estimated average variation (black
mension at baseline and the kinetics of response. The model solid line) with pointwise 95% confidence interval (dotted lines)
showed that the average ORN dimension decreased to zero and 90% prediction interval (dashed lines).

5. 836 I. J. Radiation Oncology d Biology d Physics Volume 80, Number 3, 2011
Fig. 2. Individual relative exposed bone osteoradionecrosis dimension variations in 54 patients given long-term pentox-
ifylline, vitamin E, and clodronate: observed values (o) and model-based prediction (solid lines).
(Table 3): local pain, fistula, osteitis, and trismus reductions The patient’s age, time since RT, trigger factors, previous
until disappearance. Mean SOMA scores improved signifi- treatment or type of head-neck cancer treatment had no effect
cantly (p < 0.0001): SOMA2 28 Æ14%, SOMA4 48 Æ19%, on the progression of ORN healing (NS).
SOMA6 64 Æ22%, SOMA9 73 Æ21%, SOMA12 81 Æ15%, Regular X-rays and dental computed tomography assess-
SOMA18 88 Æ13%, SOMA24 91 Æ11%, SOMA30 96 Æ5% ment showed slow but gradual and delayed improvement,
and SOMA36 98 Æ4%. with more homogeneous bone (Fig. 4).
Two-thirds (36/54) of treated patients underwent seques-
trectomy with 5-10 mm mean diameter (3-20 mm) during
DISCUSSION
the medical consultation, whereas the other 18/54 patients
did not undergo SEQ: 19/36 patients with 1 SEQ (53%) dur- The therapeutic value of PENTOCLO in ORN was first il-
ing PENTOCLO, 8 patients with 2 SEQ, 5 patients with 3 lustrated in a woman with severe 7-cm exteriorized radionec-
SEQ, and 4 patients with 4 SEQ. These 36 patients had a total rosis of the sternum, 29 years after breast cancer irradiation,
of 65 sequestrectomies in 18 months, 80% (52/65) in the first who ehibited complete healing and restoration on magnetic
6 months of treatment: 31 SEQ in (M1-M2), 13 SEQ in (M3- resonance imaging after 3 years of treatment (9).
M4), 9 SEQ in (M5-M6), 8 SEQ in (M7-M9), 3 SEQ in (M12), The present study emphasizes that refractory exposed
2 SEQ in (M18). Each SEQ preceded a level of local improve- mandible ORN is still frequent and always severe. Patients
ment then better healing, after a kind of foreign body extrac- with ORN have to be treated aggressively and quickly before
tion with purulence and foul smell. any specific treatment effect; 5 died because of fatal sepsis,

6. PENTOCLO trial in osteoradionecrosis d S. DELANIAN et al. 837
infection, simple dissolution, and osteoporosis (premature
aging), with osteocytes reaching the end of their lifespan
without replacement (20). Bone gradually becomes hypocel-
lular (fewer osteoblasts) and reduced bone matrix formation,
compensated by fibrosis.
Our ORN management was based on this pathophysio-
logic understanding, with new light shed on RIF (3). This
strategy to improve bone healing consisted of (a) reduction
of infection and purulence in irradiated bone by vigorous ini-
tial 4-week antiseptic treatment with amoxicillin-clavulanate/
ciprofloxacin, fluconazole, and methylprednisolone, allow-
ing further treatment penetration and stopping ORN worsen-
ing without danger (21); (b) marked reduction of the
microscopic RIF, sometimes combined with phenotypic
reversion of the irradiated osteoblasts, which enhance osteo-
genesis by the synergistic combination of PE; and (c) arrest of
bone destruction by inhibition of osteolysis, combined with
removal of bone sequestra with clodronate.
Fig. 3. Estimated probability for exposed bone osteoradionecrosis
complete healing by treatment with pentoxifylline, vitamin E, and
clodronate: observed values (solid line), model-based prediction PENTOCLO efficacy
for study patients (dashed line), and model-based prediction for pop- Used alone, none of the drugs included in PENTOCLO
ulation (dotted line); median at 9 months. proved able to reverse RIF or ORN. They were, however, ex-
cellent antifibrotic and antinecrotic agents (11). PTX has
whereas 5 had head-neck recurrence or a second cancer. been reported in RIF to reduce pain or trismus and improve
Moreover, in our 54 treated patients with refractory ORN, some leg functional deficits (22) and to accelerate healing
progressive instead of previous local surgery and/or HBO in radiation soft tissue necrosis (23, 24). Vitamin E seemed
treatment over a mean 17 months in 69% of patients (37/ to reduce breast RIF. By contrast, combined PE is efficient
54) led to rapid improvement and total tissue restoration after and safe in experimental (25, 26) and superficial RIF, with
PENTOCLO: half the patients recovered in 6 months, two- half RIF regression at 6 months and a two-thirds maximum
thirds in 1 year, and nearly all after 2 years (median, 9 response after a mean of 2 years (14), and in good-
months). Spontaneous sequestrectomy in 2/3 patients (with- prognosis ORN (12). Clodronate is a nonaminobisphospho-
out surgical procedure) during the first 6 months of PENTO- nate, which reduces chronic inflammation by inhibiting the
CLO seems critical because it speeds the healing process: delayed hypersensitivity granuloma response, osteoclastic re-
PENTOCLO helped separate sequestra (eliminated dead sorption due to the inhibition of osteoclast recruitment and
bone) from living bone (boosted tissue), thus allowing heal- activity, and in vitro fibroblastic proliferation, and which
ing. There was no case of programmed surgery. There was no shortens osteoclast lifespan (27, 28). Clodronate used in
difference in improvement or recovery obtained in the treated large-scale trials had unexpected effects on the viscera by
patients presenting long after irradiation, with long-term preventing metastatic diffusion, thereby underlining its pos-
ORN, with or without dental extraction. In our experience, sible effects on tissues other than bone (29). Clodronate re-
major surgery was necessary only in some salvage cases (cel- duced a case of bone marrow fibrosis with normalization of
lulitis, fracture with displacement). blood counts by stopping transfusions over an 8-month pe-
Although the clinical features of ORN have been known riod, after failure of androgen-interferon treatment (30).
for decades, its pathophysiology is poorly understood. PENTOCLO allowed rapid and definitive mucosa and skin
Descriptions of ORN tissue lesions suggest either healing in mandibular ORN patients and slow progressive
hypovascular-hypoxic or bone fibroatrophy involvement new bone formation as shown by X-rays, and computed to-
(2). The role of hypoxia was suggested by the histologic areas mography. Moreover, PENTOCLO successfully treated non-
of necrosis in severely damaged irradiated tissues (4). More- exposed thoracic or pelvic ORN and radiation-induced
over, the mandible is predisposed to ischemic radionecrosis plexitis in 90 patients (Delanian, unpublished information).
because of obliteration of the inferior alveolar artery and im- PENTOCLO safety profile. Short-term safety was good
pairment of revascularization by branches of the facial artery and did not differ from that in the placebo group in our pre-
(19). The hypothesis of RIF focuses on the defective irradi- vious randomized study. PENTOCLO long-term safety in
ated bone and the imbalance between tissue synthesis and this study was good. However, we chose not to include pa-
degradation (3). Histopathologic features of ORN are a mo- tients with active cancer because of the high healing power
saic of osteogenesis areas within extended areas of osteolysis of PE and its unknown interference with cancer. Vitamin E,
(pagetoıd appearance). Defective bone tissue is a result of
¨ usually safe (31), has been reported to be protective against
several types of degradation: osteoclastic (macrophagic) re- prostate cancer, but data are not conclusive in lung cancer
sorption, osteocytic osteolysis overwhelmed by bacterial prevention. A meta-analysis of randomized trials in

7. 838 I. J. Radiation Oncology d Biology d Physics Volume 80, Number 3, 2011
Fig. 4. Images of a 51-year-old woman with 25 Â 12 mm exposed bone osteoradionecrosis for 6 months: Subjective, Ob-
jective, Management, Analytic evaluation of injury (SOMA0 at 16) with (a) a marked bone loss on a Dentascan at baseline,
mucosal recovery after 8 months of pentoxifylline, vitamin E, and clodronate, with halving of clodronate dosage because of
diarrhea (SOMA8 at 3), symptomatic normalization at M12 (SOMA12 at 1) without radiologic change, then (b) delayed
radiologic restoration at M18 despite stoppage of treatment at M12.
cardiovascular diseases found no beneficial or adverse effect clodronate in primary breast cancer (randomized trial) in
of vitamin E on survival (32). However, another meta- primary breast cancer showed significant prevention of
analysis showed that doses of vitamin E higher than 400 osteoporosis in 268 cases and improved overall survival in
IU/day for longer than 1 year in chronic cardiovascular dis- 290 patients with bone marrow micrometastasis (39). Long
ease may increase all-cause mortality (33). Bayesian model term PENTOCLO seems to be safe.
averaging in meta-analysis showed that ‘‘vitamin E intake
is unlikely to affect mortality regardless of dose’’ (34). In vitro
studies showed pro-oxidant effects of high doses of vitamin E
CONCLUSION
that were inhibited by vitamin C: a randomized trial in 8,171
women receiving daily 600 IU vitamin E, 500 mg vitamin C, PENTOCLO effectively reduces progressive septic man-
and 50 mg beta-carotene, individually or in combination, dible ORN. The impressive and rapid clinical recovery
failed to show any difference after 9.4 years of treatment achieved suggests that theory and practice could be the basis
(35). Clodronate, which has been extensively used clinically of ORN management in the future. PENTOCLO, an
over the past 20 years, is safe. Unlike aminobisphosphonates etiology-based treatment, when combined with repeated se-
(pamidronate, zoledronate), clodronate has a significant direct questrectomy, improves prognosis from poor to good; there-
action on osteoblastic cells and increases bone formation, fore, ORN management reserves extensive surgery for
without antiangiogenic effects; the in vitro effect of clodro- salvage cases (cellulitis, fracture with displacement, exten-
nate on endothelial cells and fibroblasts is particularly mar- sive exposed bone >1 cm) All drugs are available, inexpen-
ginal (36), and no serious case of osteonecrosis of the jaw sive, well tolerated, and safe. Further randomized clinical
has yet been reported (37, 38). Three years of adjuvant trials are necessary to confirm these results.

8. PENTOCLO trial in osteoradionecrosis d S. DELANIAN et al. 839
REFERENCES
1. Balogh J, Sutherland S. Osteoradionecrosis: A review. J Otolar- 21. Annane D, Bellisant E, Bollaert P, et al. Corticosteroids in the
yngol 1989;18:245–250. treatment of severe sepsis and septic shock in adults: A system-
2. Delanian S, Lefaix J-L. Mature bone necrosis: From recent path- atic review. JAMA 2009;301:2362–2375.
ophysiological aspects to a new therapeutic action (in French). 22. Okunieff P, Augustine E, Hicks J, et al. Pentoxifylline in the
Cancer Radiother 2002;6:1–9. treatment of radiation-induced fibrosis. J Clin Oncol 2004;22:
3. Delanian S, Lefaix J-L. The radiation-induced fibro-atrophic 2207–2213.
process: Therapeutic perspective via the antioxidant pathway. 23. Dion M, Hussey D, Doornbos J, et al. Preliminary study of pen-
Radiother Oncol 2004;73:119–131. toxifylline in the treatment of late radiation soft tissue necrosis.
4. Marx R, Johnson R. Studies in the radiobiology of osteoradio- Int J Radiat Oncol Biol Phys 1990;19:401–407.
necrosis and their clinical significance. Oral Surg Oral Med 24. Futran N, Trotti A, Gwede C. Pentoxifylline in the treatment of
Oral Pathol 1987;64:379–390. radiation-related soft tissue injury: Preliminary observations.
5. Epstein J, Wong F, Stevenson-Moore P. Osteoradionecrosis: Laryngoscope 1997;107:391–395.
Clinical experience and a proposal for classification. J Oral 25. Dunn C, Galinat L, Wu H, et al. Demonstration of novel antiar-
Maxillofac Surg 1987;45:104–110. thritic and anti-inflammatory effects of diphosphonates. J Phar-
6. Epstein J, Van der Meij E, McKenzie M, et al. Post-radiation macol Exp Ther 1993;226:1691–1698.
osteonecrosis of the mandible. Oral Surg Oral Med Oral Pathol 26. Boerma M, Roberto K, Hauer-Jensen M. Prevention and treat-
1997;83:657–662. ment of functional and structural radiation-induced in the rat
7. Annane D, Depondt J, Aubert P, et al. Hyperbaric oxygen ther- heart by pentoxifylline and alpha-tocopherol. Int J Radiat Oncol
apy for radionecrosis of the jaw: A randomized placebo- Biol Phys 2008;72:170–177.
controlled, double blind, trial from the ORN96 Study Group. 27. Carano A, Teitelbaum S, Konsek J, et al. Biphosphonates di-
J Clin Oncol 2004;22:4893–4900. rectly inhibit the bone resorption activity of isolated osteoclasts
8. Ayad T, Kolb F, De Mones E, et al. The musculo-mucosal facial in vitro. J Clin Invest 1990;85:456–461.
artery flap (FAMM): Harvesting technique and indications. Ann 28. Diel I, Solomayer E, Costa S, et al. Reduction in new metastasis
Chir Plast Esthet 2008;53:487–494. in breast cancer with adjuvant clodronate treatment. N Engl J
9. Delanian S, Lefaix J-L. Complete healing of severe osteoradio- Med 1998;6:357–363.
necrosis by treatment combining pentoxifylline, tocopherol and 29. Froom P, Elmalah I, Braester A, et al. Clodronate in myelofibro-
clodronate (case report). Br J Radiol 2002;75:67–469. sis: A case report. Am J Med Sci 2002;323:115–116.
10. Ansher M. The irreversibility of radiation-induced fibrosis: Fact 30. Bendich A, Machlin L. Safety of oral intake of vitamin E. Am J
or folklore? J Clin Oncol 2005;23:8551–8552. Clin Nutr 1988;48:612–619.
11. Delanian S, Lefaix J-L. Current management for late normal tis- 31. Eidelman R, Hollar D, Hebert P, et al. Randomized trials of vi-
sue injury: radiation-induced fibrosis and necrosis. Sem Radiat tamin E in the treatment and prevention of cardiovascular dis-
Oncol 2007;17:99–107. ease. Arch Intern Med 2004;164:1552–1556.
12. Delanian S, Depondt J, Lefaix J-L. Major healing of refractory 32. Miller E, Pastor-Bariuso R, Dalal D, et al. Meta-analysis: High
mandible osteoradionecrosis after treatment combining pentox- dosage vitamin E supplementation may increase all-cause mor-
ifylline and tocopherol: A phase II trial. Head Neck 2005;27: tality. Ann Intern Med 2005;142:37–46.
114–123. 33. Berry D, Wathen J, Newell M. Bayesian model averaging in
13. Delanian S, Lefaix J-L, Maisonobe T, et al. Significant clinical meta-analysis: Vitamin E supplementation and mortality. Clin
improvement in radiation-induced lumbosacral polyradiculop- Trials 2009;6:28–41.
athy by a treatment combining pentoxifylline, tocopherol and 34. Lin J, Cook N, Albert C, et al. Vitamins C and E and beta car-
clodronate (PENTOCLO). J Neurol Sci 2008;275:164–166. otene supplementation and cancer risk: A randomized con-
14. Delanian S, Porcher R, Rudant J, et al. Kinetics of response to trolled trial. J Natl Cancer Inst 2009;101:14–23.
long term treatment combining pentoxifylline and tocopherol in 35. Walter C, Klein M, Pabst A, et al. Influence of biphosphonates
patients with superficial radiation-induced fibrosis. J Clin Oncol on endothelial cells, fibroblasts, and osteogenic cells. Clin Oral
2005;23:8570–8579. Investig 2010;14:35–41.
15. Pavy J-J, Denekamp J, Letschert J, et al. for the EORTC Late 36. Ruggiero S, Mehrotra B, Rosenberg T, et al. Osteonecrosis of
Effects Working Group. Late effects toxicity scoring: The the jaws associated with the use of biphosphonates: A review
SOMA scale. Int J Radiat Oncol Biol Phys 1995;31:1043–1047. of 63 cases. J Oral Maxillofac Surg 2004;62:527–534.
16. Schwartz G. Estimating the dimension of a model. Ann Stat 37. Diel I, Fogelman I, Al-Nawas B, et al. Pathophysiology, risk
1978;6:461–464. factors and management of biphosphonate-associated osteonec-
17. Turnbull B. The empirical distribution function with arbitrarily rosis of the jaw: Is there a diverse relationship of amino- and non
grouped, censored and truncated data. J Royal Stat Soc Series B amino-biphophonates? Crit Rev Oncol Hematol 2007;64:
1976;38:290–295. 198–207.
18. Delanian S, Porcher R, Rudant J, et al. Kinetics of response to 38. Saarto T, Vehmanen L, Blomqvist C, et al. Ten-year follow-up
long term treatment combining pentoxifylline and tocopherol in of three years of oral adjuvant clodronate therapy shows signif-
patients with superficial radiation-induced fibrosis. J Clin Oncol icant prevention of osteoporosis in early-stage breast cancer.
2005;23:8570–8579. J Clin Oncol 2008;26:4289–4295.
19. Bras J, De Jonge H, Van Merkesteyn J. Osteoradionecrosis of the 39. Diel I, Jaschke A, Solomayer E, et al. Adjuvant oral clodronate
mandible: Pathogenesis. Am J Otolaryngol 1990;11:244–250. improves the overall survival of primary breast cancer patients
20. Dambrain R. Osteoradionecrosis pathogenesis (in French). Rev with micrometastases to the bone marrow: A long term follow-
Stomatol Chir Maxillofac 1993;94:140–147. up. Ann Oncol 2008;19:2007–2011.