High or low blood pressure in patients during surgery to repair a spinal cord injury may contribute to poorer outcomes, suggests a study published today in eLife.
The findings suggest that maintaining an optimal blood pressure range during surgery may help patients recover motor function. They also show how machine learning techniques may help scientists answer key clinical questions using real-world data.
Current guidelines for treating patients with spinal cord injuries recommend avoiding low blood pressure, ensuring the adequate flow of blood to the injury site to allow recovery. But blood pressure that is too high can cause bleeding in the spinal cord that can add to the damage.
“The precise window for optimal blood pressure management to promote recovery from acute spinal cord injury is poorly understood,” explains co-first author Abel Torres-Espin, Assistant Professor of Neurological Surgery at the University of California, San Francisco (UCSF), US. “We set out to apply machine learning analytics to blood pressure and heart rate changes in operating rooms. The idea was to test the associations between these factors during surgery and neurorecovery to determine treatment thresholds that forecast recovery.”
Torres-Espin and the interdisciplinary research team analysed blood pressure data from 118 patients who underwent surgery for spinal cord injuries at the Zuckerberg San Francisco General Hospital and the Santa Clara Valley Medical Center.
“We found that patients with blood pressure that was either too high or too low during surgery for an acute spinal cord injury had poorer neuromotor recovery after surgery,” says co-first author Jenny Haefeli, Assistant Professor of Neurological Surgery at UCSF Weill Institute for Neurosciences. The team also found that patients had the best chance at recovery if their mean blood pressure was maintained between 76 mmHg and 104–117mmHg.
The results suggest that more precise upper and lower blood pressure targets may help physicians maximise their patients’ odds of recovering after a spinal cord injury, although these findings need to be confirmed by other studies. But if they are, they could lead to more use of machine learning tools to guide the care of patients with spinal cord injuries.
“Machine learning tools could be used to create real-time models that help predict the likelihood of a patient's recovery,” concludes senior author Adam Ferguson, Professor of Neurological Surgery at UCSF School of Medicine. “Such models could also be applied to forecasting patient outcomes early after a spinal cord injury.”