Risk-Prediction Model Helps Decide Use of Dialysis Prior to Ambulance Transport

October 31, 2018

By Brian Hoyle

SAN DIEGO -- October 30, 2018 -- A risk-prediction model enables emergency medical responders to determine whether urgent dialysis is necessary prior to ambulance transport, according to researchers presenting at Kidney Week 2018, the Annual Meeting of the American Society of Nephrology.

Predictors include time since last dialysis, a presenting complaint of weakness, and patient vital signs, noted presenter Karthik Tennankore, MD, MS, Dalhousie University, Halifax, Nova Scotia, Canada, at a poster session here on October 26.

Dr. Tennankore and colleagues identified 197 patients (mean age 67.2 years; 66% male, 88% white) in the region surrounding Halifax, Nova Scotia who required dialysis 3 times each week and who had been transported by ambulance to the emergency department (ED) at least once from 2009 to 2013. The study’s primary outcome was urgent dialysis, which was defined as the need for dialysis within 24 hours of admittance in either the ED, intermediate-care unit, or intensive-care unit, or dialysis that occurred within 24 hours of hospitalisation because the potassium level exceeded 6.5 mmol/L.

The patients required 624 ambulance transports to the ED, with some being transported on 2 or more different occasions. There were 87 episodes of urgent dialysis that required prompt attention.
The clinicians took into account a battery of information that included the capability of prompt transport to a facility capable of urgent dialysis, the timing of the previous haemodialysis, the nature of the main complaint at the time of paramedic assessment, a Canadian Triage and Acuity Scale score of 1 or 2, and vital signs.

The risk-prediction model offered good discrimination of cases requiring urgent dialysis, with a receiver operator curve C-statistic of 0.81 (95% confidence interval [CI]: 0.76 to 0.86). The negative predictive value for not needing urgent dialysis was 93.6% using the optimal cut point (≥15% predicted probability). Of the patients who were predicted to need urgent dialysis but who were transported to a facility that was not able to provide dialysis, one-third needed transport to another facility for urgent dialysis.

“This model has the potential to guide dialysis-patient transport to dialysis-capable facilities when needed,” Dr. Tennankore noted.

Reasons that could be discerned for urgent dialysis included weakness as the presenting complaint (odds ratio [OR] 4.62; 95% CI: 1.23 to 17.29), elapsed time over 24 hours since the previous dialysis (OR 2.09; 95% CI: 1.15 to 3.81), and triage vitals of concern monitored by paramedics. These vitals included heart rate under 60 beats/minute (OR 3.06; 95% CI: 1.09 to 8.61), systolic blood pressure over 160 mmHg (OR 2.37; 95% CI: 1.32 to 4.25), respiratory rate 20 or more breaths/minute (OR 2.00; 95% CI: 1.06 to 3.75) and oxygen saturation under 90% (OR 3.04; 95% CI: 1.55 to 5.94).

Patient comorbidities included diabetes (54%), congestive heart failure (31%), coronary-artery disease (41%), and peripheral vascular disease (26%). Causes of end-stage renal failure included diabetes (34%), glomerulonephritis (13%), ischaemic renal disease (23%), polycystic kidney disease (2%), and other/unknown causes (28%).This study was limited by the potential for data misclassification. Additionally, the model was not designed for one-size-fits-all use. It may not apply to hospitals in smaller communities or in rural settings.

Dr. Tennankore and colleagues were prompted to develop the model because patients can be in need of urgent dialysis, which often cannot be delivered until they arrive at the ED. The time spent in the ambulance on route to the hospital can be detrimental for these patients.

[Presentation title: Predicting Who Needs Urgent Dialysis Prior to Ambulance Transport to the Emergency Department. Abstract FR-PO784]