Case-control study; clinical prediction rule.Objective.
To construct and internally validate a clinical prediction rule to identify patients at high risk of developing heterotopic ossification (HO) after spinal cord injury (SCI).Summary of Background Data.
HO after SCI can lead to loss of joint mobility, loss of function, peripheral nerve entrapment, and pressure sores. HO is usually diagnosed based on clinical symptoms, as no laboratory tests are yet available to identify patients with early HO formation. Risk factors include: a complete spinal cord injury; patient age; spasticity; urinary tract infection (UTI); pneumonia; pelvic trauma; cervical or thoracic injury; and pressure sores.Methods.
We analyzed a total of 558 patients with SCI, of which 221 developed HO of the hip. A parametric survival model was fitted to estimate the probability of developing HO of the hip within three months of a spinal cord injury. Hazard ratios (HR) calculation, internal validation, calibration and model reduction were performed over 200 bootstrapped resamples. A risk score for clinical used was developed.Results.
Risk factors contributing to the risk score were completeness of the injury, age, sex, urinary tract infection, spasticity and pneumonia. The model demonstrated good discrimination (AUC = 0.72). According to the risk score quintiles, the risk of developing HO after spinal cord injury was 1.0 for persons with a score of 0 to 17 and increased 2.47-fold for persons with a score of 1 to 21, 4.75-fold for persons with a score of 22 to 27, 6.95-fold for persons with a score of 28 to 31, and 9.23-fold for persons with a score of 32-35.Conclusions.
The risk score demonstrated good discrimination in predicting the occurrence of HO within three months of a spinal cord injury. Further development and validation of the model in other populations is warranted.Conclusions.
Level of Evidence: 3