Is it possible to predict the severity of acute appendicitis? Reliability of predictive models based on easily available blood variables

Table 3 Multivariate analysis by binomial logistic regression and model fit measures

model coefficients—gangrenous appendicitis
Predictor	Estimate	SE	Z	p	Odds ratio
Intercept	0.72593	0.73657	0.9856	0.324	2.067
Albumin	− 0.07223	0.02966	− 2.4357	0.015	0.930
CRP	0.00994	0.00756	1.3148	0.189	1.010
Lymphocytes	− 1.09416	0.22605	− 4.8403	< .001	0.335
Neutrophils	− 0.41807	0.16457	− 2.5405	0.011	0.658
WBC	0.47877	0.15503	3.0882	0.002	1.614
CRP/Albumin	− 0.32336	0.11364	− 2.8455	0.004	0.724
CRP/MPV	0.05080	0.07029	0.7227	0.470	1.052
Albumin/MPV	− 0.01318	0.24328	− 0.0542	0.957	0.987
Neutr/Lymph	− 0.00422	0.01296	− 0.3256	0.745	0.996

Estimates represent the log odds of "Gangrenous Appendicitis = 1" versus "Gangrenous Appendicitis = 0"
Dependent variable: gangrenous appendicitis. Only variables that were significant at univariate analysis were introduced into the multivariate analysis. CRP: C-reactive protein, MPV: mean platelet volume and WBC: white blood cells