The use of intravenous bisphosphonates (pamidronate or zoledronic acid) is the cornerstone for the management of multiple myeloma-(MM-) related bone disease. However, osteonecrosis of the jaw (ONJ) is a rare, but sometimes difficult to manage, adverse effect of bisphosphonates therapy. A retrospective review of all MM patients who were treated with bisphosphonates in our department, from 2003 to 2013, and developed ONJ was performed. According to inclusion criteria, 38 patients were studied. All these patients were treated as conservatively as possible according to the American Association of Oral and Maxillofacial Surgeons criteria. Patients were managed with observation, oral antibacterial mouth rinse with chlorhexidine, oral antibiotics, pain control with analgesics, nonsurgical sequestrectomy with or without simultaneous administration of antibiotics, or major surgery with or without antibiotics. Healing of the lesions was achieved in 23 (60%) patients who were treated with conservative measures; the median time to healing was 12 months (95% CI: 4–21). The number of bisphosphonates infusions influenced the time to healing: the median time to healing for patients who received <16 infusions was 7 months and for those with >16 infusions was it 14 months (P = 0.017). We conclude that a primarily nonsurgical approach appears to be a successful management strategy for bisphosphonate-related ONJ.

Specific genetic polymorphisms (SNPs) have been correlated with the development of bisphosphonate-related osteonecrosis of the jaw (BRONJ) in small series. We screened 140 myeloma patients (36 patients with and 104 without BRONJ) for the presence of previously identified SNPs in PPARG and CYP2C8 genes. All the patients received exclusively zolendronic acid (ZA) therapy and were followed prospectively for BRONJ. SNPs in both genes were associated with a higher risk of development of early BRONJ, occurring within less than 2 years of ZA therapy (59% vs. 16%, p = .022 for PPARG and 29% vs. 7%, p = .07 for CYP2C8) and a shorter time to develop BRONJ (59% versus 12%, p = .011 for PPARG and 29% versus 0% at 2 years, p = .037 for CYP2C8), independently of indices of poor oral hygiene. Thus, although preliminary, our data indicate that the presence of SNPs in PPARG and CYP2C8 genes may be associated with increased risk of early BRONJ.

The administration of ZA in patients with symptomatic myeloma is associated with reduction of SREs and possible improvement of survival; its use however, is associated with increased risk of osteonecrosis of the jaw (ONJ). Longer exposure and more infusions of ZA have been associated with higher incidence of ONJ. It has been suggested that longer intervals between ZA infusions may reduce the risk of ONJ; however, this has not been proven. Since 2008, we have adopted a strategy in which ZA is discontinued in patients who remain in remission for more than 2 years, while ZA is administered every 8-12 weeks, after the first year, in patients who have achieved vgPR or CR. The aim of our analysis was to assess the incidence of ONJ in a prospectively studied population treated with ZA and to asses potential risk factors related to dosing and schedule of ZA. Since 2003, all patients in the Department of Clinical Therapeutics (Athens, Greece) undergo dental evaluation before ZA initiation and are instructed to avoid procedures that predispose to ONJ. Only patients who received ZA and survived at least 6 months post their first ZA infusion were included in the analysis. Relative dose frequency of ZA (RDF) was calculated as the average number of weeks between infusions of ZA (weeks between first and last infusion divided by the number of infusions). Time of exposure to ZA was calculated from the date of first infusion until the date of last infusion. Death due to myeloma and development of ONJ were treated as competing events. Between January 2000 and January 2013, 266 patients with symptomatic myeloma fulfilled the above criteria and were included in the analysis. Median age was 68 years (range 36-87 years) and 47% were males. Median follow up was 36 months and 35% of the patients have died. The median number of ZA infusions was 16 (range 1-107) and the median time of exposure to ZA was 29 months, corresponding to 9073 person-months of exposure. Median RDF was one infusion per 7.9 weeks (interquartile range (IQR) 5.8-11.4 weeks). ONJ developed in 26 (10%) patients. Median time from first ZA infusion to development of ONJ was 28 months (range 6-122 months). Accounting for death as a competing event, 1-year risk of ONJ was 1.2% (95% CI 0.3-3%), 2-year risk was 5% (95% CI 2.5-8.5%), 3-year risk was 8.5% (95% CI 5-13%), 4-year risk was 11.5% (95% CI 7.3-17%) and 5-year risk was 14% (95% CI 9-19%) (Figure). The respective survival rates were 93%, 84%, 76%, 65% and 56% at 1, 2, 3, 4 & 5 years. The median time of exposure to ZA was 28 months (IQR 22-46 months) for patients who developed ONJ vs. 24 months (IQR 11-42 months) for those who did not (p=0.2). However, the median number of ZA infusions was 23 for those who developed vs. 14 for those who did not develop ONJ (p=0.004). Increasing number of ZA infusions was associated with higher risk of ONJ: 2.3% of patients who received <10 infusions developed ONJ vs. 12% of patients who received 10-19 infusions vs. 15% of patients who received 20 or more ZA infusions (p=0.012). The incidence rate of ONJ for patients who started ZA before 2008 was 0.31 per 100 person-months vs. 0.26 per 100 person-months for those who started after 2008 (p=0.4). The incidence of ONJ at 3-years was 13.6% (95% CI 8-20%) for patients with an RDF <8 weeks vs. 2.6% (95% CI 0.5-8%) for patients with RDF of ≥8 weeks (p=0.018). However, after adjusting for RDF of ZA infusions, only the number of ZA infusions remained significant (p=0.03). The multivariate analysis showed that both the number and the frequency of ZA infusions were associated with a shorter time to ONJ development. More specifically, average frequency of infusion <8 weeks was associated with a 15-fold (95% CI 4-55, p<0.001) increase in the risk of ONJ, while for every infusion of ZA the risk of ONJ increased by 9% (95% CI 4-14%, p<0.001). In conclusion, our data, in a large population of consecutive unselected patients with symptomatic myeloma who received ZA, indicate that ONJ remains a frequent complication in patients who receive ZA for prolonged periods and the risk increases with the number of ZA infusions. More frequent infusions of ZA are associated with earlier development of ONJ, but the risk of ONJ is associated mainly with the cumulative dose of ZA. Prospective clinical trials should examine if less frequent administration of ZA can reduce the risk of ONJ without compromising its antiresorptive effect. Disclosures: No relevant conflicts of interest to declare.

Bisphosphonates are extensively used in the treatment of myeloma-related bone disease as they reduce pain and skeletal related events (SREs), such as pathologic fractures, need for radiation and surgery to the bone. One of the complications of prolonged therapy with bisphosphonates, especially of amino-bisphosphonates, such as zoledronic acid (ZA) is osteonecrosis of the jaw (ONJ). Recent studies have attempted to correlate specific genetic polymorphisms (SNPs) and ONJ. Such SNPs include cytochrome P4502C8 (CYP2C8; Sarasquete et al, Blood 2008) and peroxisome proliferator-activated receptor gamma (PPAR-γ; Di Martin et al, Br J Haematol 2011) that have been linked to increased risk of ONJ in two different series. The aim of this study was to investigate a possible association between SNPs in CYP2C8 and PPAR-γ and the risk of developing ONJ in a large number of MM patients who received ZA. We screened 36 patients who developed ONJ and 104 patients who did not develop ONJ for the SNPs of interest in PPAR-γ (rs1152003) and CYP2C8 (rs193495) genes by direct sequencing of peripheral blood derived DNA. All patients were treated with similar systemic anti-myeloma therapies during the same period of time, had received only ZA as antiresorptive therapy and were followed prospectively for the development of ONJ in a single center (Department of Clinical Therapeutics, University of Athens, Greece). The median follow up of the cohort was 72 months. The median number of ZA infusions was 27 (range: 4-107). Patients who developed ONJ had a median of 31 infusions versus 25 infusions for patients who did not develop ONJ. However, 31% of patients who developed ONJ had received less than 24 infusions of ZA. Median time to development of ONJ was 47 months (range: 7-182 months). The median relative dose intensity for ZA was 2.85 mg/month; it was 3.3 mg/month for those who developed ONJ and 2.7 mg/month for those who did not. An extraction preceded the development of ONJ in 60% of patients who developed ONJ, it was unprovoked in 20%, it was associated with trauma from dentures in another 15% and in 5% ONJ was preceded by a periodontal and/or periappical inflammation (abscess formation etc). There was no significant difference in frequency of the presence of SNPs in the two studied genes between patients with and without ONJ. However, there is a subset of patients who develop ONJ rather early, after the infusion of relatively few doses of ZA while others develop ONJ after protracted exposure to ZA. Since the most important factor associated with the development of ONJ after ZA therapy in patients with myeloma is the total dose of ZA, we analyzed separately patients who had less than 24 versus those who had at least 24 infusions of ZA. In patients with <24 infusions of ZA, the presence of SNPs in both PPAR-γ and CYP2C8 was associated with a significantly higher probability and a shorter time to development of ONJ. More specifically, the presence of SNPs in the PPAR-γ was associated with a significantly higher risk of development of ONJ with less than 24 ZA infusions (55% versus 16%, p=0.011) and at a significantly shorter time to development of ONJ (19 versus 69 months, p<0.001). The presence of SNPs in the CYP2C8 was associated with a higher risk of developing ONJ with less than 24 ZA infusions (29% versus 7%, p=0.07) and at a shorter time (44% versus 13% at 3 years, p=0.037). Combining the genotype risk, those with high risk of SNPs in both genes had a 70% cumulative incidence of ONJ within 24 months from initiation of ZA versus 17% for those carrying one of the two SNPs and 0% for those without any high risk of SNPs (p<0.001). In conclusion, our data indicate that SNPs in the CYP2C8 and PPAR-γ genes are associated with a risk of early development of ONJ. However, increasing cumulative dose of ZA increases substantially the risk of ONJ in all patients, independently of genotype-defined risk. Disclosures No relevant conflicts of interest to declare.