High frequency of osteonecrosis of the jaw among denosumab-treated prostate cancer patients

To the Editor,

Bisphosphonates and denosumab, also referred to as bone-targeted therapies, decrease the risk of skeletal-related events (SREs) in men with castration-resistant prostate cancer (CRPC) and bone metastases. Monthly zoledronic acid infusion decreases the risk of SREs as compared to placebo (38% vs. 49%) in men with CRPC [1]. In patients with castration-sensitive disease, however, zoledronic acid does not reduce the risk of SREs [2,3]. Denosumab, a monoclonal antibody against RANK ligand, further reduces the risk of SREs and increases the time to first bony complication by four months when compared to zoledronic acid [4].

Osteonecrosis of the jaw (ONJ) was first suspected as a complication of zoledronic acid and pamidronate treatment in 2003 [5,6]. Several bisphosphonates and denosumab seem to be associated with this complication, indicating a relation to the use of bone antiresorptive treatment [7]. A recent update of the pivotal trial comparing the two antiresorptives reported ONJ in 8.2% of CRPC patients with only denosumab exposure and in 5.9% of those having switched from zoledronic acid to denosumab [8].

After the initial positive results of superiority of denosumab over zoledronic acid in preventing SREs, many CRPC patients at our institution were switched from zoledronic acid to denosumab. During recent years, we have been alarmed by the high number of ONJ cases among prostate cancer patients.

The current case series includes all of the 254 men with CRPC and bone metastases who were treated at the Comprehensive Cancer Center (CCC) of Helsinki University Central Hospital (HUCH) during 2014. All CRPC patients who had received at least one dose of zoledronic acid or denosumab for their metastatic cancer were included.

The medical charts of oncology, urology, surgery, dentistry and oral or maxillofacial surgery were reviewed. Details of patient history and treatment of prostate cancer were collected with a data cutoff date of 30 June 2015. The diagnosis of ONJ was verified by an oral/maxillofacial surgeon or in few cases by a dentist. The study was accepted by the HUCH ethical committee.

In total, 29 of the 254 men (11.4%) with CRPC and bone metastases developed ONJ during bone-targeted therapy. Of the patients who developed ONJ, two had received only zoledronic acid (7%), nine only denosumab (31%) and 18 patients (62%) had received both zoledronic acid and denosumab sequentially during the course of their disease. All but four of the 29 ONJ cases (86%) were detected during denosumab treatment.

Jaw necrosis was diagnosed after a median of 19 doses of bone-targeted therapy. The two patients with only zoledronic acid exposure developed ONJ after eight doses. Nine patients with only denosumab exposure were diagnosed with ONJ after a median of 18 doses (range 6–29). The 18 patients who had received both zoledronic acid and denosumab sequentially developed ONJ after a median of 23 doses of bone-targeted therapy (range 8–52).

Dental extractions had been performed before or during bone-targeted therapy in 62% (18/29) of the patients who later developed jaw necrosis. Denosumab and zoledronic acid had been withheld for a variable time period after dental surgery. Most of the patients had a history of prolonged (≥3 months) course of corticosteroids. A total of 52% (15/29) of the patients were on corticosteroids, most often in combination with abiraterone acetate, at the time of the ONJ diagnosis. Fifty-nine percent (17/29) of the ONJ patients had received the first dose of bone modifying agent during the castration-sensitive stage of their prostate cancer.

Management of ONJ was either conservative consisting of follow-up and use of antibiotics if purulent discharge was observed or it contained surgical intervention if considered necessary. The management and spectrum of ONJ patients is presented in more detail in Table 1. Majority of the patients had grade I and II osteonecrosis or they fluctuated between the classes [9].

Table 1. Staging, grading and treatment of the osteonecrosis of the jaw patients.

Patient	Stage	Grade	Location, the most severe area first	Oral surgery prior onset	Removable dental prosthesis	Type of treatment	Number of operations
1	N/A	N/A	N/A	Yes	N/A	N/A	N/A
2	II	1A	Posterior mandible	No	Yes	Follow-up/topical CHX	0
3	I	1B	Posterior maxilla	Yes	Yes	Debridement	1
4	I	2A	Posterior mandible	No	No	Follow-up/topical CHX	0
5	N/A	N/A	N/A	N/A	N/A	N/A	N/A
6	I	3B	Posterior mandible, anterior maxilla	No	No	Debridement	1
7	II	1B	Posterior mandible, anterior maxilla	Yes	No	Curettage	1
8	I	1B	Anterior mandible	Yes	Yes	Debridement	1
9	II	2A	Posterior mandible	No	No	Debridement	2
10	II	3B	Posterior mandible, anterior mandible	Yes	No	Curettage	2
11	I	2B	Posterior mandible	Yes	No	Debridement	4
12	III	4B	Whole maxilla	Yes	No	Resection of hemi-maxilla	1
13	II	3B	Whole maxilla, whole mandible	Yes	Yes	Debridement	2
14	I	1A	Posterior mandible	Yes	Yes	Debridement	1
15	II	2B	Whole mandible	N/A	No	Debridement	2
16	N/A	N/A	N/A	N/A	N/A	N/A	N/A
17	II	4A	Whole maxilla	N/A	N/A	Debridement	N/A
18	II	1A	Posterior mandible	No	No	Debridement	1
19	II	3B	Whole mandible	No	Yes	Debridement	1
20	I	3A	Posterior mandible	Yes	No	Follow-up/topical CHX	0
21	I	3B	Posterior maxilla, posterior mandible	Yes	No	Debridement	3
22	I	3A	Posterior mandible	No	No	Follow-up/topical CHX	0
23	III	3B	Posterior maxilla, anterior mandible	Yes	N/A	Debridement/curettage	3
24	II	4A	Posterior mandible	Yes	No	Debridement	3
25	II	3B	Posterior mandible	Yes	No	Debridement	3
26	II	1B	Anterior mandible	Yes	Yes	Follow-up/topical CHX	0
27	I	3B	Posterior mandible, posterior maxilla	Yes	No	Debridement	1
28	II	3B	Posterior mandible	Yes	N/A	Debridement, incision	3
29	II	2B	Whole mandible, anterior maxilla	Yes	N/A	Debridement	2

The staging of ONJ is based on clinical findings as presented in [9]. I: Exposed/necrotic bone in patients who are asymptomatic and have no evidence of infection; II: Exposed/necrotic bone associated with infection as evidenced by pain and erythema in the region of the exposed bone with or without purulent drainage; III: Exposed/necrotic bone in patients with pain, infection, and one or more of the following: pathologic fracture, extra-oral fistula or osteolysis extending to the inferior border or sinus floor or mandible. The grading of ONJ lesions is based on size and number of exposed bone areas: A means single lesion and B multiple lesions; 1: lesions under 5 mm in diameter, 2: largest lesion under 10 mm in diameter, 3: largest lesion under 20 mm in diameter and 4: largest lesion over 20 mm in diameter. Anatomical location of the ONJ is considered posterior if located behind canine region and anterior if located in between canines. CHX: chlorhexidine-based mouth rinse; N/A: Not available, as some of the patients with ONJ were treated also outside the Helsinki University Central Hospital Maxillofacial Surgery Unit.

In total, 59% (17/29) of the ONJ cases had chronic disability or jaw necrosis continued until death of the patient. Of the 17 patients with permanent disability, three had received only denosumab and 14 patients both denosumab and zoledronic acid sequentially. Thirty-one percent (9/29) of the ONJ cases showed marked improvement. Of the nine patients showing improvement of jaw necrosis, three had received only denosumab, one only zoledronic acid and five both denosumab and zoledronic acid sequentially. Data on the outcome of ONJ was missing in three patients.

In our case series, ONJ was noted in 11.4% of CRPC patients treated with bone antiresorptives. In the literature, ONJ has been reported in 1–6.5% of prostate cancer patients after zoledronic acid exposure of variable duration [4,10–12]. The risk of denosumab-associated ONJ in randomized studies of prostate cancer patients is 2–8.2% [4,8,13]. With all bone-targeted therapy, the risk of jaw necrosis rises with higher cumulative doses and longer exposure [12].

Only one third of our ONJ cases showed marked improvement of the condition whereas the rest had chronic disability. This is in line with data from the pivotal denosumab versus zoledronic acid trial, where 26% of the prostate cancer patients had resolution of jaw necrosis [8]. Thus, ONJ is often a chronic condition subjecting patients to numerous surgical debridement operations and long periods of antibiotic therapy.

Obvious limitations of our study include its retrospective case series design and heterogeneous patient population. Our institution is a referral unit and many of the CRPC patients had received bone antiresorptives with various protocols already before attending the CCC department of HUCH. Our current practice is to avoid bone-targeted therapy in prostate cancer patients with castration-sensitive disease, but many of the patients in this ‘real world series’ had received bone antiresorptives early on. The frequency of ONJ was likely underestimated. Asymptomatic cases of jaw necrosis have been missed as formal monitoring of oral health was not incorporated in our routine oncologic follow-up. The cumulative doses of denosumab and zoledronic acid were checked from medical charts but data was not conclusive.

A recent meta-analysis suggested that the risk of denosumab-associated ONJ is three-fold higher in patients with prostate cancer as compared to those with other cancer types [14]. The risk factors of osteonecrosis such as poor oral hygiene, corticosteroids and anemia are common among prostate cancer patients. Unlike bone-targeted therapy, modern anticancer treatments with docetaxel, cabazitaxel, abiraterone acetate, enzalutamide and Ra 223 increase survival of CRPC patients [15]. Although men with CRPC live longer they also receive higher cumulative doses of bone antiresorptives.

Most patients in the current case series had been given both bone-targeted therapies sequentially. A recent update of the pivotal trial comparing the two antiresorptives reported ONJ in 8.2% of CRPC patients with only denosumab exposure and in 5.9% of those having switched from zoledronic acid to denosumab [8]. Thus, switching from one bone-targeted treatment to another does not seem to increase the risk of ONJ, but data on this subject is very limited. Although the sequential use of bone antiresorptives does not seem to increase the risk of ONJ, prognosis of the condition seems better if only denosumab has been used. This may be due to differences in the metabolism of the two agents: zoledronic acid retains in the skeleton for years while denosumab does not incorporate in bone [8,12,16].

Recently, a study on 1822 patients with bone metastases from multiple myeloma, breast cancer or prostate cancer compared zoledronic acid administered every three months to the standard monthly dosing. The less frequent dosing was shown to be non-inferior as the risk of SREs was similar during the three-monthly and monthly dosing of zoledronic acid. ONJ developed twice as often during the monthly as the three-monthly dosing [17]. Comprehensive data on the less intensive dosing of denosumab is lacking. REDUCE is a large phase III study currently recruiting prostate and breast cancer patients with bone metastases which randomizes patients to monthly or three-monthly dosing of denosumab.

The risk of medication associated ONJ during bone-targeted therapies seems higher in clinical practice as compared to that reported in randomized trials. Our single institution case series found an 11.4% frequency of ONJ among prostate cancer patients. Only one third of the ONJ cases showed marked improvement whereas the rest caused chronic disability. ONJ is a common and often chronic complication of denosumab and zoledronic acid and more effective preventative measures are warranted.

Disclosure statement

Leena Vehmanen and Tapio Utriainen: scientific congress costs and lecture fees from pharmaceutical companies. Carl Blomqvist, Risto Kontio and Juho Suojanen: participation in an Amgen study project.