Predictors of delayed failure of structural kyphoplasty for pathological compression fractures in cancer patients.

OBJECT Pathological compression fractures in cancer patients cause significant pain and disability. Spinal metastases affect quality of life near the end of life and may require multiple procedures, including medical palliative care and open surgical decompression and fixation. An increasingly popular minimally invasive technique to treat metastatic instabilities is kyphoplasty. Even though it may alleviate pain due to pathological fractures, it may fail. However, delayed kyphoplasty failures with retropulsed cement and neural element compression have not been well reported. Such failures necessitate open surgical decompression and stabilization, and cement inserted during the kyphoplasty complicates salvage surgeries in patients with a disease-burdened spine. The authors sought to examine the incidence of delayed failure of structural kyphoplasty in a series of cement augmentations for pathological compression fractures. The goal was to identify risk predictors by analyzing patient and disease characteristics to reduce kyphoplasty failure and to prevent excessive surgical procedures at the end of life. METHODS The authors retrospectively reviewed the records of all patients with metastatic cancer from 2010 to 2013 who had undergone a procedure involving cement augmentation for a pathological compression fracture at their institution. The authors examined the characteristics of the patients, diseases, and radiographic fractures. RESULTS In total, 37 patients underwent cement augmentation in 75 spinal levels during 45 surgeries. Four patients had delayed structural kyphoplasty failure necessitating surgical decompression and fusion. The mean time to kyphoplasty failure was 2.88 ± 1.24 months. The mean loss of vertebral body height was 16% in the patients in whom kyphoplasty failed and 32% in patients in whom kyphoplasty did not fail. No posterior intraoperative cement extravasation was observed in the patients in whom kyphoplasty had failed. The mean spinal instability neoplastic score was 10.8 in the patients in whom kyphoplasty failed and 10.1 in those in whom kyphoplasty did not fail. Approximately 50% of the kyphoplasty failures occurred at junctional spinal levels. All the patients in whom kyphoplasty failed had fractures in 3 or more cortical walls before treatment, whereas 46% of patients in the nonfailure group had fractures with breaching of 3 or more walls. CONCLUSIONS Although rare, delayed failures of structural augmentation with cement during kyphoplasty do occur and can lead to additional surgeries. A possible predictive index may include wall integrity of the vertebral body, competency of the posterior tension band, and location of the kyphoplasty at a junctional spinal level. Additional studies are required to confirm these findings.