This study estimated the cumulative incidence and risk ratio for osteonecrosis of the jaw (ONJ) after tooth extraction in patients with and without administration of bisphosphonates (BP) and identified potential risk factors for bisphosphonate-induced osteonecrosis of the jaw (BIONJ). A cohort study was conducted in all patients undergoing tooth extraction at a university hospital in Japan from April 2006 to June 2009. Of 3216 patients, 126 had BP administration, of whom 5 (3.9%, 95% confidence interval (CI): 1.2–9.2) developed ONJ, versus 1 (0.032%, 95% CI: 0.00081–0.18) among 3090 patients without BP administration. BP administration was associated with the development of ONJ after tooth extraction, with an unadjusted risk ratio of 122.6 (95% CI: 14.4–1041.8). When stratified by age and route of BP administration, the risk ratio for ONJ patients aged 65 years or older with intravenous BP administration compared to those without was 200.2 (95% CI: 23.8–1679.4, P < 0.001). Patients receiving BP showed a significant association between the incidence of BIONJ and alveolar bone loss score. The risk of ONJ is higher in patients with than without BP administration, particularly intravenous administration. Severe periodontitis might be a risk factor for BIONJ.
Bisphosphonates (BP) are used for the treatment of a range of bone problems, such as osteoporosis or bone metastases of malignant cancer, and their efficacy in preventing further bone damage, reducing bone pain, and increasing bone mineral density has been confirmed. In 2003, Marx reported bisphosphonate-related osteonecrosis of the jaw or bisphosphonate-induced osteonecrosis of the jaw (BIONJ), as a side effect of BP treatment. The incidence and mechanism of BIONJ have not been determined accurately.
Since that initial report, the association between BP exposure and the incidence of osteonecrosis of the jaw (ONJ) has been clarified in several case series, reviews, epidemiologic studies and clinical trials, which reported a prevalence of BIONJ ranging from 0.7% to 18.6% on intravenous and 0.01% to 4.3% on oral administration. The low incidence of ONJ among BP-naïve patients has prevented any direct estimation of the risk ratio of ONJ among patients receiving BP. Black et al. identified only 1 patient with possible ONJ among 3852 postmenopausal women during a 3-year period. In their 6-year population-based cohort study using medical claims data, Wilkinson et al. found that 0.30% of naïve patients had been diagnosed with inflammatory conditions or osteomyelitis of the jaw but not ONJ.
Tooth extraction has been reported to be the main initiating factor and one of the most common risk factors for BIONJ among patients receiving BP (approximately 86% of cases), and relative risk of BIONJ in these patients is 5.3–53 times higher than in BP patients who do not experience tooth extraction. A current guideline recommends non-surgical treatment rather than tooth extraction in dental patients at high risk of BIONJ, but given that bacterial infection is reported as a critical risk factor for BIONJ, avoiding extraction might be problematic in cases in which the bacterial infection remains. Information on the incidence or risk factors for BIONJ after tooth extraction among patients receiving BP is limited. A better understanding of this condition, particularly with regard to risk factors and incidence, will be helpful to dentists in the care of patients receiving treatment with BP.
The purpose of this study was to estimate the cumulative incidence and risk ratio for ONJ after tooth extraction in patients with and without administration of BP, and to identify potential risk factors for BIONJ, including oral status.
Materials and methods
The authors conducted a retrospective analysis of patients who had undergone tooth extraction between April 2006 and June 2009 at the Department of Oral and Maxillofacial Surgery, Kyoto University Hospital. Patients were identified using administrative data, and dental and medical records were reviewed from January 2010 to August 2010. This study protocol was approved by the Ethics Committee of Kyoto University Graduate School and Faculty of Medicine.
Inclusion was restricted to patients aged 20 years or older. ONJ in this study was diagnosed by the presence of exposed bone in the maxillofacial region that had persisted for more than 8 weeks, in reference to diagnostic criteria formulated by the American Association of Oral and Maxillofacial Surgeons. Diagnoses were determined independently by two oral and maxillofacial surgeons. Among patients undergoing tooth extraction at the authors’ institution, the following were excluded: tooth extraction in patients with metastatic tumours to the oral region; signs of osteomyelitis or exposed bone in the maxillofacial region before extraction; any history of craniofacial radiation therapy; and patients for whom panoramic radiographs before extraction were not available. Cases with diagnostic disagreement were discussed. Finally, ONJ cases with diagnostic disagreement were regarded as non-ONJ cases and analysed with the non-ONJ patients. The kappa value for inter-observer agreement was 0.86 (95% confidence interval (CI); 0.82–0.89).
For each chart reviewed, the following information was collected: demographics; medical history; test results; and potential risk factors associated with ONJ, namely the use of steroids, chemotherapy (including anticancer agents, immunosuppressive agents and thalidomide), current smoking status, current alcohol intake, diabetes, and details of BP treatment (indication, type, dose, duration). Recommended regimens for oral BP administration in the treatment of osteoporosis in Japan are etidronate sodium, 200 mg per day (Didronel ® ); alendronate sodium, 5 mg per day or 35 mg per week (Fosamac ® or Bonalon ® ); and risedronate sodium, 2.5 mg per day or 17.5 mg per week (Actonel ® or Benet ® ), while those for intravenous administration in the treatment of osteolytic bone metastases of malignant cancer, multiple myeloma, or hypercalcemia of malignancy are incadronate, 10 mg per time (Bisphonal ® ); pamidronate, 30–45 mg for hypercalcemia of malignancy, or 90 mg every 4 weeks for osteolytic bone metastases of malignant cancer (Aredia ® ); and zoledronic acid, 4 mg every 4 weeks (Zometa ® ). Intravenous BP has not been approved for osteoporosis in Japan.
For patients treated with BP in the hospital, these medications were entered into the electronic medical record system to obtain the type and duration of BP administration, and the number of patients to whom they were administered, to estimate the incidence of BIONJ. For patients treated with BP in other hospitals, the authors reviewed the record of their first examination at the authors’ hospital, at which the type and duration of BP administration as obtained from the referring physician by letter is recorded.
Using the patient’s panoramic radiograph before tooth extraction, an experienced examiner calculated the DMF index and severity of alveolar bone loss. The DMF index, which comprises the number of decayed (D), missing (M), and filled (F) teeth, has been established as a key measurement of caries experience in dental epidemiology. The severity of alveolar bone loss was measured to examine periodontal status as a percentage of missing bone at the mesial and distal surfaces of each tooth present. Severity was estimated from ( a − b )/ a × 100 (%) using the panoramic radiograph (where a is the distance from radiographic apex to cement–enamel junction (mm) and b is the distance from radiographic apex to interproximal alveolar bone crest (mm)). If the location of the cemento-enamel junction was obscured by interproximal fillings, the cervical margin of these fillings was chosen as standard. If the cervical margin of these fillings was obscured, bone loss height was characterized as unmeasurable. Each tooth surface was assigned a score corresponding to a bone loss of 0–24%, 25–49%, 50–74%, and 75–100%, respectively. These measurements were averaged to yield a single mean bone loss score for each patient, with a higher score indicating more severe periodontal disease.
The incidence of ONJ was calculated using the cumulative incidence method, which is defined as the number or proportion of a cohort of people who experience the onset of ONJ during a specified time interval. In the calculation of the CI for incidence, the Poisson distribution was used. Medians for continuous variables were compared using the Wilcoxon rank-sum test. Proportions across levels of categorical variables were compared using the Fisher exact test. Risk ratios were calculated for all dichotomous variables and Wald CIs were calculated. All P values were two sided at a significance level of 5%. For missing data, available-case analysis was performed in addition to multiple imputation analysis. All statistical analysis was performed using Stata 11.2 software (Stata Corporation, College Station, TX, USA).
3240 patients underwent tooth extraction at the authors’ institute. After exclusion of 8 patients with a history of craniofacial radiation therapy and 16 without a panoramic radiograph, 3216 (99.2%) eligible patients were entered into the analysis. Patient characteristics are summarized in Table 1 . 126 patients (3.9%) had received BP; among the 103 women (81.7%), 86 (83.5%) were treated with oral BP and 17 (16.5%) with intravenous BP, while among the 23 men, 13 (56.5%) received oral and 10 (43.5%) received intravenous BP. Seven patients were diagnosed with both osteoporosis and breast cancer, all of whom had been treated with oral BP. The total number of patients without BP administration was 3090 (96.0%), of whom 51.8% were women. Median age was lower (39.0 years) than that of patients administered BP (66.0 years).
|Yes ( n = 126)||No ( n = 3090)|
|Male||23 (18.3)||1490 (48.2)|
|Female||103 (81.7)||1600 (51.8)|
|Primary disease (%)|
|Osteoporosis *||99 (78.6)||74 (2.4)|
|Breast cancer *||13 (10.3)||61 (2.0)|
|Multiple myeloma||11 (8.7)||5 (0.2)|
|Prostate cancer||1 (0.8)||41 (1.3)|
|Kidney cancer||3 (2.4)||1 (0.03)|
|Other cancer||6 (4.8)||184 (5.6)|
|Route of BP administration (%)|
Cumulative incidence of ONJ after tooth extraction
Five of the 126 patients receiving BP developed ONJ (3.9%, 95% CI: 1.2–9.2), versus 1 (0.032%, 95% CI: 0.00081–0.18) of the 3090 without BP administration. Individual data from the 6 patients with ONJ are shown in Table 2 . The cumulative incidence of ONJ among patients with BP administration was significantly higher than that among those without ( P < 0.001). The crude risk ratio for ONJ after tooth extraction for patients with BP administration compared to those without was 122.6 (95% CI: 14.4–1041.8). By route of administration, cumulative incidence was 1.0% (95% CI: 0.025–5.6) among patients treated with oral BP and 14.8% (95% CI: 4.0–37.9) in those treated with intravenous BP. The risk ratio for ONJ was 31.2 (95% CI: 1.9–495.4) among patients treated with oral BP and 457.7 (95% CI: 52.8–3962.7) in those treated with intravenous BP. Large differences were seen in age and prevalence of cancer or osteoporosis between BP and BP-naïve patients. The authors therefore performed stratified analysis by age and route of BP administration to estimate risk ratios for ONJ to control for these factors ( Table 3 ). Among patients aged 65 years or older (median 73, range 65–94 years), the risk ratio for ONJ for patients with oral or intravenous BP administration compared to those without was 12.9 (95% CI: 0.82–204.6, P = 0.138) or 200.2 (95% CI: 23.8–1679.4, P < 0.001), respectively.
|Sex||Age||Primary disease||Type of BP||Dosage of BP (mg)||Duration of BP (months)||Other risk factors||Cure and prognosis|
|F||75||Breast cancer||P||1455||45||Chemo therapy||Currettage, nonhealing|
|M||75||Prostate cancer||Z||32||8||Chemo therapy||Die of primary disease|
|M||67||Kidney cancer||Z||28||7||None||Sequestrectomy, healing|
|ONJ (+) ( n = 6)||ONJ (−) ( n = 3210)||Total ( n = 3216)||Risk ratio (95% CI)||P value|
|BP (+)||5 (83.3)||121 (3.8)||126||122.6 (14.4–1041.8)||<0.001|
|BP (−)||1 (16.7)||3089 (96.2)||3090||1.0 (ref)|
|Oral BP (+)||1 (16.7)||98 (3.1)||99||31.2 (1.9–495.4)||0.061|
|IV BP (+)||4 (66.6)||23 (0.7)||27||457.7 (52.8–3962.7)||<0.001|
|BP (−)||1 (16.7)||3089 (96.2)||3090||1.0 (ref)|
|Age-stratified (≥65 years)|
|BP (+)||5 (16.7)||63 (8.3)||68||51.5 (6.1–434.8)||<0.001|
|BP (−)||1 (83.3)||700 (91.7)||701||1.0 (ref)|
|Age-stratified (<65 years)|
|BP (+)||0||58 (2.4)||58||N.A.||N.A.|
|BP (−)||0||2389 (97.6)||2389||1.0 (ref)|
|BP- and age-stratified (≥65 years)|
|Oral BP (+)||1 (16.7)||53 (7.0)||54||12.9 (0.82–204.6)||0.138|
|IV BP (+)||4 (66.6)||10 (1.3)||14||200.2 (23.8–1679.4)||<0.001|
|BP (−)||1 (16.7)||700 (91.7)||701||1.0 (ref)|
|BP- and age-stratified (<65 years)|
|Oral BP (+)||0||45 (1.8)||45||N.A.||N.A.|
|IV BP (+)||0||13 (0.53)||13||N.A.||N.A.|
|BP (−)||0||2389 (97.6)||2389||1.0 (ref)|
Association with possible risk factors for BIONJ after tooth extraction
The authors investigated potential risk factors associated with BIONJ among patients with BP administration. The incidence of BIONJ was significantly associated with the type of BP ( P = 0.007) but not with the other potential risk factors ( Table 4 ). The crude risk ratio for intravenous compared with oral BP administration was 14.6 (95% CI: 1.7–125.8). The type, dosage and duration of intravenous BP administration are shown in Table 5 . Median duration in patients with BIONJ (13.5, range 7–49 months) was longer than that in patients without BIONJ (7.5, range 1–96 months), but this difference was not significant. Regarding oral BP, etidronate, alendronate and risedronate were given as single agents to 4 (4.0%), 54 (54.5%), and 31 patients (31.3%), respectively, while the rest (10.2%) received a series of double BP agents.
|Variable||Bisphosphonate-induced osteonecrosis of the jaw|
|Yes ( n = 5)||No ( n = 121)||RR (95% CI)||P value|
|Male||2 (40.0)||21 (17.4)||2.9 (0.52–16.8)||0.225|
|Female||3 (60.0)||100 (82.6)||1.0 (ref)|
|Administration route—no. (%)|
|Intravenous||4 (80.0)||23 (19.0)||14.6 (1.7–125.8)||0.007|
|Oral||1 (20.0)||98 (81.0)||1.0 (ref)|
|Steroid use—no. (%)|
|Yes||2 (40.0)||54 (44.6)||0.83 (0.14–4.8)||1.000|
|No||3 (60.0)||67 (55.4)||1.0 (ref)|
|Chemotherapy use—no. (%)|
|Yes||3 (60.0)||23 (19.0)||5.7 (1.0–32.7)||0.059|
|No||2 (40.0)||98 (81.0)||1.0 (ref)|
|Yes||0 (0)||9 (7.4)||N.A.||N.A.|
|No||5 (100)||112 (92.6)|
|Current smoking—no. (%) *|
|Yes||0 (0)||21 (18.4)||N.A.||N.A.|
|No||5 (100)||93 (81.6)|
|Current alcohol intake—no. (%) *|
|Yes||1 (20.0)||29 (25.4)||0.74 (0.086–6.3)||1.000|
|No||4 (80.0)||85 (74.6)||1.0 (ref)|
|Bone loss score|