Positive predictive value of the International Classification of Diseases, 10th revision, codes to identify osteonecrosis of the jaw in patients with cancer

Background: Osteonecrosis of the jaw (ONJ) is an important adverse event associated with therapies suppressing bone turnover, especially in patients with high-dose regimens of antiresorptive therapy, such as cancer patients. Danish health registries are an important resource for monitoring side effects of drugs. The International Classification of Diseases, 10th revision (ICD-10), currently used in Denmark, does not have a specific code for ONJ, making it difficult to monitor its occurrence. Objectives: To estimate the positive predictive value (PPV) for ONJ of currently used ICD-10 codes, suggested by Danish oral and maxillofacial surgeons, in order to assess feasibility of identification of ONJ cases among cancer patients in the Danish National Registry of Patients (DNRP). Methods: This study was conducted in northern Denmark (1.8 million inhabitants) among patients with a history of cancer. In Denmark ONJ cases are referred to hospital-based departments of oral and maxillofacial surgery (DOMS). In the DNRP, we identified patients with potential ONJ diagnosed at DOMS (as suggested by a series of ICD-10 codes) from 1 January 2005 to 31 December 2009. To confirm or rule out ONJ, we reviewed hospital records of these patients originating from DOMS. A confirmed ONJ case was defined by the presence of exposed maxillofacial bone for 8 weeks or more, in the absence of previous craniofacial radiation therapy. The PPV was the proportion of confirmed cases among all potential cases. Results: Among 85,910 eligible cancer patients, we identified 91 (0.11%) potential cases of ONJ, of which 18 were confirmed. The overall PPV was 20% (95% CI: 12-29%), ranging from 0% to50% for individual ICD-10 codes. Conclusions: A majority of cases identified by the suggested ICD-10 codes did not fulfill the criteria for ONJ, even though the potential cases were identified at DOMS. Therefore, reliance on ICD-10 codes, without hospital chart review, will lead to an overestimation of the occurrence of ONJ among cancer patients.