Bevacizumab-Associated Implant Presence-Triggered Osteonecrosis: A Case Report and Literature Review

Bevacizumab is a recombinant humanized monoclonal antibody designed to selectively bind with and inhibit the biological activity of all human vascular endothelial growth factor inhibitor isoforms. It is mainly used for treating advanced cancers such as metastatic colon, kidney, brain, lung, and breast cancer.¹  In 2008, Greater et al²  described a case of osteonecrosis of the jaw (ONJ) after tooth extraction as a side effect of bevacizumab therapy; this particular patient had not received bisphosphonates, and bevacizumab was determined to have caused the ONJ. Subsequently, the American Association of Oral and Maxillofacial Surgeons, in its 2014 position paper, renamed bisphosphonate-related osteonecrosis of the jaw (BRONJ) as medication-related osteonecrosis of the jaw (MRONJ), which includes osteonecrosis associated with other antiabsorption (denosumab) and angiogenesis-inhibiting (sunitinib, bevacizumab, among others) therapies.³ 

The risk factors for developing MRONJ include operative treatment,^4–6  anatomic factors,^4–8  concomitant oral disease,^5,7–11  and demographic, systemic, and other medication-related factors.^4–6,8–10  The placement of dental implants, along with surgical procedures for tooth extraction and periodontal and root cysts, is a known risk factor for MRONJ. However, implant-related MRONJs have been observed after administration of bone-modifying agents (BMAs) or angiogenesis inhibition therapy in the presence of an implant. Thus, patients with a pre-existing dental implant, receiving bisphosphonates, may develop ONJ.¹²  Giovannacci et al¹³  called these conditions implant presence-triggered osteonecrosis (IPTO).

However, there is still a lack of consensus regarding the administration of bevacizumab therapy to patients having dental implants. The purpose of this manuscript is to report a rare case of IPTO related to bevacizumab therapy and to review the relevant literature.

A 66-year-old woman complained of repeated episodes of swelling and pain in relation to the buccal gingiva of implants at the #19 and #20 positions since February 2020. This mild to moderate pain and gingival swelling developed spontaneously and subsided within a few days. By July 2020, the patient was unable to close her mouth properly because of an uneven bite. The patient visited a local dental clinic, where on clinical examination, there was exudate in relation to the implants at #19 and #20 (Brånemark MkII, placed 8 years ago). Panoramic radiography showed pathologic fractures of the mandible resulting from osteolysis and isolation involving the regions corresponding to positions #19 and #20 (Figure 1a). The failed implants at the #19 and #20 positions were removed along with the peri-implant necrotic bone en bloc with a tweezer (Figure 1b). However, residual necrotic bone was observed in the surrounding region. The patient subsequently visited our department for the treatment of osteomyelitis. Written informed consent was obtained from the patient for the publication of this case report and the accompanying images.

When the patient first visited us, her left cheek appeared red and swollen, with hypoesthesia of the left part of the lower lip. Her mouth opening was restricted (1.5 finger breadth), and the occlusion was displaced to the left (Figure 2a). There were multiple implants present in her oral cavity (in relation to #3, #4, #5, #7, #11, #12, #13, #14, #21, #22, #23, #24, #25, #27, and #29). Her oral hygiene maintenance was poor, and the probing depth was >5 mm around all implants. In addition, several implants showed exposed threads (#3, #4, #5, #11, #12, #13, #21, #22, #23, #24, and #25). Necrotic bone and purulent discharge were observed in the regions corresponding to #19 and #20 (Figure 2a). There was mandibular displacement because of a pathologic fracture, and the left mandibular region was in contact with the occlusal surface of #14. This premature contact caused the patient pain and led to an increase in the necrosed area of the mandible.

A panoramic radiograph revealed a pathologic fracture in relation to #19 and #20, extending from the alveolar crest to the inferior border of the mandible (Figure 2b). Metastatic bone tumor was excluded on histologic examination, and osteomyelitis was suspected. Thus, a diagnosis of MRONJ (IPTO) because of bevacizumab therapy (stage 3) was established. Microbiological examination revealed Prevotella denticola and Prevotella intermedia as the causative organisms of the MRONJ.

The patient had been diagnosed with breast cancer with metastases (T1N3M1) in September 2018. The patient had no other known comorbidities and no history of smoking. Because the patient had metastatic triple-negative breast cancer and was considered incapable of undergoing radical surgery, chemotherapy had been advised. The patient had no history of head and neck radiation therapy or administration of BMAs. The patient had received bevacizumab (10 mg/kg every 2 weeks) and paclitaxel (90 mg/m² at cycles D1 and D15). The infusion was initiated in October 2018, and 19 kur (38 bevacizumab infusions) had been administered by the time the patient visited us.

We discussed the treatment for ONJ with the patient and the patient's oncologist. Because of the poor prognosis of metastatic breast cancer and the need for chemotherapy for the metastases, the patient opted out of definitive mandibular resection and reconstructive surgery. In addition, we reached a consensus for the suspension of bevacizumab and paclitaxel therapy until the inflammation had subsided. Plaque and calculus were removed, and the patient was given oral hygiene instructions. The screw-retained prosthesis on #14 was also removed. The patient visited the Center for Infectious Diseases, and antibacterial agents, including amoxicillin (1500 mg/day) + clavulanic acid (375 mg/day), were initiated and continued for 4 weeks. Although the acute inflammatory symptoms were alleviated, prolonged diarrhea was noted as a side effect. The medication was changed to amoxicillin (750 mg/day) + clindamycin (900 mg/day) and the diarrhea disappeared. By August 2020, there was no pain or purulent discharge. With the ceasing of the acute inflammation, bevacizumab and paclitaxel therapies were resumed.

In September 2020, sequestrum separation was observed around the left lower molar region (#19 and #20 region). Removal of the sequestrum was performed accompanied by systemic administration of antibacterial agents (Figure 3a). Underneath the removed necrotic bone was juvenile epithelial tissue. Histopathologic examination of the removed necrotic bone ruled out the possibility of tumor metastases to the jaw (Figure 3b). Fourteen days after debridement, the left lower molar region was completely covered with soft tissue with no bone exposure, and the patient was no longer in pain (Figure 4a through c).

Mandibular defect reconstruction was suggested as the choice of treatment; however, the patient refused any further surgical intervention because of poor prognosis of the breast cancer. At the time of her first visit to us, her diet consisted of pureed foods. Even after the acute inflammation had disappeared, her occlusion was greatly altered because of the pathologic fractures. Masticatory efficiency was measured using GLUCO SENSOR GS-II (GC Co Ltd, Tokyo, Japan), and the glucose concentration obtained from a chewed gummy jelly was 69 mg/dL. A splint with twin rows on the defect side was fabricated based on the patient's centric occlusion relation (Figure 5); this increased her dietary intake. The value of glucose concentration on the GLUCO SENSOR GS-II increased to 142 mg/dL.

No signs of bone dissolution or isolation were observed on the panoramic radiographs and computed tomography, which showed regeneration at the inferior border of the mandible on the left side (Figures 6a and 7). Amoxicillin (750 mg/day) + clindamycin (900 mg/day) was discontinued 4 months after conservative surgery. Currently, at 6-month follow-up after conservative surgery, the wound surface of the resected area revealed healthy mucosal coverage without lesions. The patient continued bevacizumab and paclitaxel therapy until December 2020, which was then changed to adriamycin and cyclophosphamide therapy because of the progression of breast cancer. However, MRONJ treatment helped to maintain sufficient mouth opening and dietary intake.

Dental implants have been proven to improve the oral health and function of patients. However, as with teeth, cases of ONJ have been reported around implants and have become a major concern among clinicians.^14,15  Lazarovici et al¹⁶  were the first to study cases of BRONJ associated with implants. They classified BRONJ occurring after at least 6 months of implant placement as spontaneous osteonecrosis and stated that patients receiving bisphosphonate therapy who undergo implant placement need a long-term follow-up to detect the onset of implant-related BRONJ. Kwon et al¹⁷  coined the term “implant surgery-triggered osteonecrosis” to define osteonecrosis occurring within the first 6 months of implant placement. In their series of 19 patients, 58% of cases with osteonecrosis were not associated with surgery, and 3 cases occurred in patients who underwent implant placement before bisphosphonate administration. Troeltzsch et al¹²  conducted a retrospective study on 36 patients with dental implants who received antiresorptive therapy with functional loading of the implants from 2010 to 2016. They observed that MRONJ may be strongly associated with peri-implantitis. A study by Giovannacci et al¹³  describes 15 patients with peri-implant osteonecrosis. The patients were categorized as having “implant surgery-triggered” MRONJ (ISTO) if the osteonecrosis occurred immediately after implant placement (2−10 months). Patients who developed MRONJ much later after implant placement (1–15 years) were defined as having “implant presence-triggered” MRONJ. There were 9 cases of IPTO, 8 of which occurred in patients receiving intravenous treatment with zoledronic acid for tumors. Escobedo et al¹⁸  reviewed 74 cases of BMA-related IPTO and 27 cases of ISTO. They noted that most cases of peri-implant MRONJ were unrelated to implant placement (alveolar surgery) and had developed in relation to previously osseointegrated implants.

Kwon et al¹⁷  have characterized the bone destruction patterns of BRONJ lesions around implants into 3 types: frozen, osteolytic, and en bloc. Of these, the en bloc type, which maintains direct bone contact with the dental implant, may be associated with microcracks in the mandible owing to the mechanical stress during mastication.¹⁹  It is known that dental infections are involved in the development of MRONJ. Bacterial infection is thought to occur when exposed necrotic bone infiltrates around the implant.¹⁷  This pattern could result from local bone destruction around the implant and severe infection and could exist simultaneously, depending on the degree of local bone destruction and the severity of the infection. IPTO results in a large number of pathologic fractures of the mandible, similar to the findings in the present case,¹⁸  in which we observed en bloc type of destruction and pathologic fractures in the mandible.

The exact cause of ONJ associated with antiangiogenic drugs is not well understood. Theoretically, angiogenesis inhibition adversely affects the bone regeneration ability after bone invasion, delays remodeling and healing, and increases susceptibility to superinfection.^20,21  Additionally, it is known to inhibit macrophage chemotaxis and osteoblast differentiation.²²  MRONJ occurs in approximately 1%−10.6% of patients receiving BMA therapy during the advanced stages of cancer.^23–26  In contrast, the incidence of ONJ because of bevacizumab was 0.3%−0.4% in 2 double-blind, randomized trials (AVADO and RIBBON-1) and in a large, nonrandomized safety study (ATHENA) on advanced breast cancer. Additionally, the incidence of ONJ in bisphosphonate-exposed patients receiving bevacizumab was 0.9%−2.4%.^27–29  The case reports of these 3 clinical trials did not include a detailed assessment of oral health and were limited to the toxicity assessment according to the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) 3.0.³⁰  Nonetheless, the incidence of bevacizumab-related ONJ is considered to be much lower than that with BMA therapy. This report describes a single case, and the pathophysiology and treatment of IPTO related to bevacizumab therapy require further elucidation. Reporting of more such cases could improve our understanding of this kind of lesion in the future.

In this study, articles published until March 2021 were searched in the database Medline via PubMed. We searched for articles containing the following keywords: “MRONJ” AND “implant” AND “bevacizumab.” Of the 2 case reports that satisfied our search criteria, 1 report was excluded because it described a case of IPTO in a patient who received bevacizumab in addition to BMA. Therefore, there were only 2 cases of bevacizumab-related ONJ occurring around implants, including the present case (Table 1).³¹  There was 1 case each of ISTO and IPTO. In both cases, bevacizumab therapy was indicated for breast cancer with metastasis, with a total of 32 infusions for the patient with ISTO and 38 infusions for the patient with IPTO. In the 2 cases of bevacizumab-related osteonecrosis of the jaw, maintenance and improvement of oral health, antibacterial mouthwash, systemic administration of antibacterial agents, and cessation of bevacizumab therapy were performed to treat MRONJ. In both cases, debridement or removal of sequestra was performed locally after observing bone isolation, and subsequent mucosal covering of the lesion was seen. Bevacizumab therapy was resumed in both cases. A healthy mucosal covering without lesions was seen with no sign of recurrence. In addition, the follow-up radiographs showed no sign of bone lysis or sequestration.

Different disciplines have reported perioperative wound complications in patients receiving bevacizumab therapy. Past reviews have concluded that performing surgery 5–6 weeks after cessation of bevacizumab therapy reduced the rate of bevacizumab-related perioperative wound complications.^32,33  This may be related to the estimated half-life of bevacizumab of approximately 20 days (range, 11–50 days). Theoretically, extensive surgical treatment is more beneficial in treating MRONJ than conservative surgery. However, in advanced MRONJ with a tumor-bearing state, radical treatment does not always improve the patient's quality of life. In these cases, administration of antibacterial agents, cessation of bevacizumab therapy for 5–6 weeks, and conservative surgery may be preferable.

In conclusion, to the best of our knowledge, we have reported the first case of bevacizumab-induced IPTO in a patient in the absence of BMA treatment. In recent years, the number of patients receiving bevacizumab and BMA treatment has increased. Clinicians should recognize MRONJ as a potential side effect of bevacizumab treatment. This often occurs after tooth extraction or oral surgery, but it can also be unrelated to them. IPTO associated with bevacizumab is an uncommon but potentially severe disease for patients using functional implants. Appropriate dental interventions and frequent oral management during the course of the treatment may prevent IPTO in patients receiving bevacizumab.

Abbreviations

BMAs:

bone-modifying agents

BRONJ:

bisphosphonate-related osteonecrosis of the jaw

IPTO:

implant presence-triggered osteonecrosis

ISTO:

implant surgery-triggered osteonecrosis

MRONJ:

medication-related osteonecrosis of the jaw

ONJ:

osteonecrosis of the jaw

The authors are grateful to Takashi Miyazaki from Miyazaki Dental Clinic for detailed discussions that were instrumental in our case report. The authors also thank Editage (www.editage.jp) for English language review.

The authors declare no conflicts of interest.