Dynamic prediction for clinically relevant pancreatic fistula: a novel prediction model for laparoscopic pancreaticoduodenectomy

Table 2 Univariate analysis of drainage fluid amylase and C-reactive protein to albumin ratio in postoperative days

Landmark time point	Variables	Patients without CR-POPF	Patients with CR-POPF	P value
POD1	Drainage fluid amylase, IU/L	1224 (195–2482)	1082 (700–2143)	0.533
POD1	CAR	3.16 (2.31–4.19)	3.15 (0.52–7.23)	0.916
POD2	Drainage fluid amylase, IU/L	630 (137––2013)	1149 (653–2651)	0.041
POD2	CAR	4.84 (3.27–6.17)	6.36 (3.71–8.30)	0.055
POD3	Drainage fluid amylase, IU/L	248 (56–934)	978 (415–4990)	< 0.001
POD3	CAR	4.57 (3.12––6.14)	6.44 (4.20–9.51)	0.024
POD5	Drainage fluid amylase, IU/L	63 (20–240)	454 (178–3604)	< 0.001
POD5	CAR	2.62 (1.37–3.72)	3.78 (2.46–5.79)	0.027

Variables with P < 0.10 (in bold) were to be included in the next multivariate analysis
POD postoperative day, CAR C-reactive protein to albumin ratio, AUC area under curve, ROC receiver operator characteristic curve

ISSN: 1471-2482