The specific characteristics and independent predictors of no-reflow phenomenon, development of a clinically-adaptable risk estimation system

Abstract Introduction No-reflow (NR) phenomenon occurs, when myocardial perfusion is not re-established despite opening the coronary artery during percutaneous coronary intervention (PCI). Purpose Our aim was to identify no-reflow specific characteristics, its independent predictors, and to develop a clinically-adaptable risk score. Methods We have analysed 4085 patient data from two Hungarian cardiovascular centres. We included all STEMI/NSTEMI patients underwent PCI (n=3187). 158 patients treated with papaverine/adenosine formed NR group, while 3029 patients were in control (C) group. Anamnestic parameters, laboratory and operation data were compared. Statistical analysis was carried out with Mann-Whitney-, Fisher test, binary logistic regression and Kaplan Meier survival curve. Based on our results, we designed a risk estimation system, checking its applicability with ROC analysis. Results As for NR-specific characteristics, malignant ventricular arrhythmias (11% vs. 4%, p=0,0031; NR-C consequently) and complications (21% vs. 11%, p=0,064) showed their vulnerability. The increment of glucose (8,1 vs. 7,1 mmol/l, p=0,004), WBC (12,08 vs. 10,5 G/l, p=0,001), CRP (12,46 vs. 7,67 mg/l, p=0,051) and LDL levels (3,34 vs. 3,13 mmol/l, p=0,059) supported the pathomechanism of NR. Higher biomarker levels (troponinT: 2040 vs. 510,5 ng/ml; CK-MB: 100,4 vs. 63,65 U/l, p<0,0001) indicated severe perfusion disturbance. Tendency was seen in higher BMI (28,65 vs. 28,03 kg/m2, p=0,12). STEMI dominated in NR (83 vs. 59%, p<0,0001). Lower platelet level (213,3 vs. 228 G/l, p=0,107) and single vessel disease (46 vs. 25%, p=0,0042) characterized NR. 30-day survival was significantly different (85,1 vs. 93,54%, p<0,0001). The mortality rate of NR in STEMI was 69,7% (69,7% vs. 7,94%, p<0,0001) and in NSTEMI 3,7% (3,7% vs. 4,32%). From the significant differences, CRP was the independent predictor of NR (OR: 1,011, p=0,004; pro 1 mg/l change). Examining STEMI/NSTEMI separately, in STEMI CRP was the independent predictor (OR: 1,0092, p=0,036). In NSTEMI LDL (OR: 4,23, p=0,021) was the independent factor. In the risk score, the following 8 parameters were included: BMI>28 kg/m2, glucose>8 mmol/l, WBC>12 G/l, CK-MB>100 U/l, hs troponin T>2000 ng/ml, CRP>12 mg/l, LDL>3,3 mmol/l, STEMI (yes/no), thus maximum 8 points could be reached. Low (0–1 points, 5–20%), moderate (2–5 points, 55–70%) and high risk groups (6–8 points, 41–11%) were formed. Supervising the model with ROC analysis: AUC=0,69, p=0,0026, which indicates its ability to discriminate effectively between different risk levels of NR. Conclusions The specific characteristics of NR group were identified, from which CRP was the independent predictor - as well as in STEMI, while in NR-NSTEMI LDL was the independent factor. With the elaborated risk estimation system –using anamnestic and routine laboratory parameters– NR could be predicted and unsuccessful PCI could be reduced, resulting in positive therapeutic consequences.