Perfusion index: Could this be a new triage tool for upper gastrointestinal system bleeding in the emergency department? A prospective cohort study

ABSTRACT BACKGROUND: Many scoring systems for predicting mortality, rebleeding and transfusion needs among patients with upper gastrointestinal bleeding (UGIB) have been developed. However, no scoring system can predict all these outcomes. OBJECTIVE: To show whether the perfusion index (PI), compared with the Rockall score (RS), helps predict transfusion needs and prognoses among patients presenting with UGIB in emergency departments. In this way, critical patients with transfusion needs can be identified at an early stage. DESIGN AND SETTING: Prospective cohort study in an emergency department in Turkey, conducted between June 2018 and June 2019. METHODS: Patients’ demographic parameters, PI, RS, transfusion needs and prognosis were recorded. RESULTS: A total of 219 patients were included. Blood transfusion was performed in 174 patients (79.4%). The PI cutoff value for prediction of the need for blood transfusion was 1.17, and the RS cutoff value was 5. The area under the curve (AUC) value for PI (AUC: 0.772; 95% confidence interval, CI: 0.705-0.838; P < 0.001) was higher than for RS (AUC: 0.648; 95% CI: 0.554-0.741; P = 0.002). 185 patients (84.5%) were discharged, and 34 patients (15.5%) died. The PI cutoff value for predicting mortality was 1.1, and the RS cutoff value was 7. The AUC value for PI (AUC: 0.743; 95% CI: 0.649-0.837; P < 0.001) was higher than for RS (AUC: 0.725; 95% CI: 0.639-0.811; P < 0.001). CONCLUSION: PI values for patients admitted to emergency departments with UGIB on admission can help predict their need for transfusion and mortality risk.

研究背景：目前已开发出多种用于预测上消化道出血（upper gastrointestinal bleeding, UGIB）患者死亡率、再出血风险及输血需求的评分系统，但尚无单一评分系统可同时覆盖上述全部结局指标。研究目的：探讨与罗卡爾評分（Rockall score, RS）相比，灌注指数（perfusion index, PI）是否有助于急诊科室收治的上消化道出血患者预测其输血需求及预后，以期早期识别存在输血需求的危重患者。研究设计与研究场所：本研究为前瞻性队列研究，于2018年6月至2019年6月在土耳其某急诊科室开展。研究方法：记录患者的人口统计学参数、灌注指数、罗卡爾評分、输血需求及预后情况。研究结果：本研究共纳入219例患者，其中174例（79.4%）接受了输血治疗。预测输血需求的灌注指数截断值为1.17，罗卡爾評分截断值为5；灌注指数的受试者工作特征曲线下面积（area under the curve, AUC）为0.772，95%置信区间（confidence interval, CI）为0.705~0.838（P < 0.001），显著高于罗卡爾評分的AUC值（0.648，95%CI：0.554~0.741，P = 0.002）。本研究中185例（84.5%）患者出院，34例（15.5%）患者死亡。预测死亡率的灌注指数截断值为1.1，罗卡爾評分截断值为7；灌注指数的AUC值为0.743（95%CI：0.649~0.837，P < 0.001），略高于罗卡爾評分的AUC值（0.725，95%CI：0.639~0.811，P < 0.001）。研究结论：急诊科室收治的上消化道出血患者入院时的灌注指数值，可辅助预测其输血需求及死亡风险。